Exploratory Study of RFID Applications for Air Cargo Operations

EXPLORATORY STUDY OF RFID APPLICATIONS FOR AIR CARGO OPERATIONS


By

MAGALIE LANIEL















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2010

































2010 Magalie Laniel



























To all who guided me throughout my lifetime, making this milestone possible









ACKNOWLEDGMENTS

This work has been carried out under the guidance of my advisor Dr. Jean-Pierre

Emond. His brilliant ideas and suggestions more than helped shape this research. I will

never be thankful enough for the opportunity and support that he gave me through all

my studies. I feel very privileged and fortunate to have him as my advisor and want to

thank him for believing in me.

I would also like to thank each of my committee members, Dr. Ray Bucklin, Dr.

Tom Burks, Dr. Daniel Engels and Dr. David Mikolaitis for their advice and contribution

to my education thus far and in the future. Besides, I would like to express my

thankfulness to Ismail Uysal who helped me extensively structure this dissertation and

with whom I hope to have the pleasure to work with in the future.

I would like to send a special thank to Trevor Howard and all of Air Canada Cargo

for their generous implication and participation to this research. Without their help, many

experiments would have simply been impossible to achieve. Moreover, I want to send a

warm recognition to everyone at Franwell for helping in any way they could and

providing key equipment at critical times. Thank you for your trust and encouragement. I

also want to thank Mr. Chris Noel from Sealed Air Corporation for providing temperature

sensors in a timely manner and making my last experiment possible.

I would like to extend my gratitude towards my parents for their unconditional love

and support regardless of distance and sacrifices. I would also like to thank my other

half Martin for his patience, guidance and never ending revisions. Last but not least I

would like to thank all my friends and colleagues for their support. Particularly Cecilia, it

was such a pleasure to share this path together.









TABLE OF CONTENTS


page

A C KNO W LEDG M ENTS ............ ................ ........................... ............... 4

L IS T O F T A B L E S ........................ .................................................................................. 9

LIST OF FIGURES.................................. ......... 11

LIST O F A BBR EV IATIO NS ............ ................ ............................... ............... 15

A BSTRACT ........................ ............................................. 16

CHAPTER

1 GENERAL INTRODUCTION ............... ....................... .............. ............... 18

2 RFID IN AIR CARGO: A LITTERATURE REVIEW .............................. ............... 24

Introduction .. .... ... ......... ........ ..... ........ ................................... 24
RFID Technology and Definitions ....................................................................... 24
Radio Waves ......................... ........ .... ......... 25
Polarization of electromagnetic waves.................... ............... 26
Electromagnetic waves properties .................... ............... 27
RFID System Overview .... .. ................................ ... .... ............ 30
Readers ............................ ........ ..... ............... 30
Antenna ............... .. ............ ......... ............... 31
T a g s .............. ................. ............................................. ............... 3 3
Frequencies .......................... ......... ......... 36
Air Cargo............................................. ............... 38
Aviation History ................................... ......... 38
Air Cargo Supply Chain ..................... ....................... 39
Air cargo warehouse operations .................................. .. ..................... 40
Unit load device (ULD) ................... ..... ........................................ 41
M a rket ....... ... .. .... ..... ........ ................................................... 4 2
Materials in Commercial Aircrafts ....... ...................... ........ 43
Composites ........................ ........... ......... 43
Metals ................ ....... .................. 44
Electrical Systems in Commercial Aircrafts ....... ........ ..... .................. 45
Temperature Profile in Commercial Aircraft................... ............. 46
Aircraft Safety ........................... ........ ........ 46
Cargo security and monitoring ........................... .......... 46
Fire detection ...... ...................... .................. 47
Technologies..................... ... .................. ............... 48
RFID in Aviation ......... ............ ........ ...................... .. ....... 49
ISM Frequency and Aviation RFID Considerations ...................................... 49









Aviation Applications ............... .... ......... ............... 51
Passenger baggage sortation .............. ..................... ............. 51
Verification / authentication ................ ...... .... ....... ............... 52
Tracking and locating ...... ............. ....................... 53
Cargo ............... ........... ............... ....... ...... ......... 54
Cold chain ......... ........... ........... .... ............... 56
Wireless Interference ............................................... 58
Electronic devices ................. ............... ..... ............... 60
RFID interference ............ ................. ........... 60
RFID airworthiness policy ........ .......... .......... ............ 61
C including R em arks ........................ .................. .. .. ............................... 62

3 AIR CARGO WAREHOUSE ENVIRONMENT AND RF INTERFERENCE ............. 68

Introduction ............................. ............... 68
Materials and Methods........................................... ............... 72
M ontreal W warehouse ...................... ....... ......... .. .. ........................... 74
Toronto W arehouse.............................. ............... 74
Results and Discussion.......................................... ............... 74
433M Hz.............................. ............ ............... 75
915MHz ........... ........... .................................. 76
2.45GHz ....................... ........... .......... ............... 77
C o nclusio n ............. ......... .. .............. .. .. .................. ............... 7 8

4 RADIO FREQUENCY PROPAGATION INSIDE THE CARGO HOLD OF A DC-
10 AIRCRAFT........................... ......... 86

Introduction .......................................... 86
Materials and Methods........................................... ............... 89
Test 1: Propagation Study ...... ............. .................... ..... 90
Data analysis .. ................................. .......................................... 90
Statistical analysis...................... ............................. ................. 94
Test 2: Validation of Relation between Signal Strength and Tag Reads ....... 95
Data point comparison ...... ........... .................. ......... 96
Results and Discussion.......................................... ............... 96
Test 1: Propagation Study ....................... ............ ......... 96
Attenuation ...................... .... ........................... 96
S signal strength ........... .. ......... ...................................... ......... .... 99
Test 2: Validation of Relation between Signal Strength and Tag Reads ........ 100
C o nclusio n ............. ......... .. .............. .. .. .................. ............... 102

5 RADIO FREQUENCY INTERACTIONS WITH AIR CARGO CONTAINER
MATERIALS FOR REAL-TIME MONITORING............................... 115

In tro d u ctio n ............. ......... .. ................................. ................. ............... 1 15
Materials and Methods......................................... 118
T e s t 1 ............. ......... .. .............. .. .................................................. 1 2 0


6









T e s t 2 .............. ..... ............ ................. ............................................ 1 2 0
T e s t 3 ........... ...................................................................... 1 2 1
Results and Discussion....................................... 122
T e s t 1 .............. .... ............. ................. ............................................ 1 2 2
433MHz................ ......... ......... ......... 122
915MHz ...................................................................... ......... ................... 123
2.45G Hz..... ......................................................................................... 123
T e s t 2 ................................................................................................ 1 2 5
T e s t 3 ................................................................................................ 1 2 5
C o n c lu s io n .............................................................................. ............... 12 7

6 TEMPERATURE MAPPING INSIDE AIR CARGO CONTAINERS DURING
AIRSIDE OPERATIONS ............... ........................... 131

In tro d u ctio n .............................................................................. ............... 13 1
Materials and Methods......................................... 134
Results and Discussion....................................... 136
D during Flight ......... .. ................ ......... .......................................... 136
Before and After Flight ............................... .................... 138
C o n c lu s io n .............................................................................. ............... 13 9

7 GLOBAL TRACKING SYSTEM FOR AIR CARGO SUPPLY CHAIN ............... 145

Introduction .................................................................................. ......... 145
Typical Air Cargo Warehouse Operations ...................................... 146
Typical Air Cargo Ramp Operations ........... ..... ........... ................ ........... 148
Findings from this Study and Recommendations for RFID Tracking System
Implementation .............. ........ ...................... 149
Passive and Active RFID Tags ............... .... ......... .......... ... ........... 150
ULD M aterials............................................... 152
Frequency ................ ......... .................. 152
C o n c lu s io n .............................................................................. ............... 15 4

8 GENERAL CONCLUSION ................ .............................. 160

APPENDIX

A DC-10 CARGO HOLD AND CARGO DOOR SPECS ............................. 162

B RADIO FREQUENCY ATTENUATION SURFACE PLOTS ............................. 164

C STATISTICAL ANALYSIS RESULTS FOR DC-10 RADIO FREQUENCY
PROPAGATION ................ ......... ......... ......... 173

D RADIO FREQUENCY SIGNAL STRENGTH PROPAGATION SURFACE
P L O T S ..................................................................................................... 1 7 9

E UNIT LOAD DEVICES, TEMPERATURE GRAPHS AND POSITIONS ........ 188









LIST OF REFERENCES .............. ..... ................. ......... 194

BIOGRAPHICAL SKETCH ............... ..... ......... .......... .......... 205









LIST OF TABLES


Table page

2-1 E P C global tag class structure .................................................. .... .. ............... 63

2-2 Permissible field strengths for RFID systems in accordance with FCC Part 15
(F C C 2 0 0 8 ) ................ ................. ........................................................ .............. 6 3

2-3 A aircraft radio system s ....................................... ............ ........ ............... 63

3-1 Receiving antenna specifications ................................ ...................... ............. 80

3-2 Minimum and maximum interference readings (in dBm) for six positions and
three frequencies at the Montreal warehouse................................................... 81

3-3 Minimum and maximum interference readings (in dBm) for nine positions and
three frequencies at the Toronto warehouse. ................................................ 81

4-1 Specifications of the three RF systems used .............. ........... ........... .... 104

4-2 Calculated parameters for the attenuation equation (Eq. 4-5) and maximum
allowed output power adjustment. .......... ............................. ............. 104

4-3 Averages and standard deviations of attenuation levels for each test ............ 104

4-4 Statistical analysis results for the effect of antenna location............................. 105

4-5 Statistical analysis results for the effect of antenna polarization....................... 105

4-6 Signal strength data for each test, averaged per vertical slice, and total cargo
hold (Avg). .................................. ....................................... ......... 105

4-7 Table summarizing the recorded and adjusted power levels and read rates
for circular and linear antennas across the 12 cross sectional planes ............. 106

5-1 Specifications of the six antennas used................... ................ ....... ........... 128

5-2 Signal strength measurements (dBm) for control (no sample), test 1.
Receiver antenna positions are measured from the emitting antenna.......... 128

5-3 Signal strength deviation, test 1. Receiver antenna positions are measured
from the emitting antenna and sample materials are positioned at ............... 128

5-4 Signal strength measurements, plus signal strength deviation between
material samples and control at three frequencies for test 2 ......................... 129

5-5 Signal strength measurement, test 3. Receiver antenna positions are
measured from the emitting antenna. ........ .......... .. .................... 129









5-6 Signal strength deviation, test 3. Receiver antenna positions are measured
from the emitting antenna and sample materials are positioned at L ............... 129

6-1 Routes, aircraft and ULD specs from Toronto (YYZ) ..................................... 140

6-2 Temperature comparison between heated and unheated cargo holds inside
an A irbus 330. ................................................................ .. ......... 140

7-1 Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo acceptance part) ..... .................. .................. 155

7-2 Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo build-up part). ... .. ...................... ............. 156

7-3 Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo to/from the ramp section)............... .... ................ 157

7-4 Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo break-down and storage section) ................ ............... 158

7-5 Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo delivery part)............................... .............. 159

C-1 Effects of frequency, antenna location and antenna polarization on
attenuation levels of the complete dataset. ....... ...... ..... .................. 173

C-2 Effect of width on attenuation levels for each frequency, antenna location and
polarization. ............. ...... ... ... ................ .......................... ......... 174

C-3 Effect of height on attenuation levels for each frequency, antenna location
and polarization. .......... ......... ... ......... ................ .... ............ 175

C-4 Effect of depth on attenuation levels for each frequency, antenna location
and polarization. ........................ .. ............................................ 176









LIST OF FIGURES


Figure page

1-1 Brief overview of the air cargo supply chain. ..... ........ ..... ................. 22

1-2 Some problems associated to the air cargo warehouse and aircraft
operations ..................... .... .......... ...... ............................... 22

1-3 Some questions associated with warehouse and aircraft air cargo operations. 23

2-1 Electrom magnetic spectrum ................................................................................... 64

2-2 Different parts of a w ave (Lahiri, 2006)................................... ..................... 65

2-3 Example of an RFID system on conveyor belt............................ ... ............... 65

2-4 Wave propagation for linear and circular polarization ............. ............... 65

2-5 Cargo warehouse floor plan and activity areas........................... ... ............... 66

2-6 International flight temperature profile with both high-temperature excursions
(during stopovers) and low-temperature excursions (in flight).......................... 67

2-7 Example of a GSM/GPS capable RFID system for real time data acquisition
(Schm oetzer, 2005) ................................................................... ............ 67

3-1 Montreal cargo warehouse facility floor plan and interference reading points
(numbered 1 to 6). .......... ......... ... ......... ............... ............... 82

3-2 Toronto cargo warehouse facility floor plan and interference reading points
(numbered 1 to 9). .......... ......... ... ......... ............... ............... 83

3-3 Noise floor of spectrum analyzer at three different resolution bandwidths ........ 83

3-4 Worst case scenario for signal interference readings around 433MHz. Span:
10MHz, RBW : 10kHz, attenuation: OdB, gain: OdBi ............. .... ...... ........... 84

3-5 Worst case scenario for signal interference readings around 915MHz. Span:
50MHz, RBW: 10kHz, attenuation: OdB, gain: 2.5dBi........................... 84

3-6 Worst case scenario for signal interference readings around 2450MHz. Span:
50MHz, RBW : 10kHz, attenuation: OdB, gain: 8dBi....................................... 85

4-1 Section of an aircraft fuselage (Airbus A380) ................................................ 106

4-2 DC-10-30F from Arrow Cargo................................. ............... 107









4-3 Typical fuselage section of a DC-10-30F, lower cargo hold circled in blue
(B oeing, 20 10c) ................................................................ ... ............ 107

4-4 Cargo hold dimensions and RF emitting antenna positions ........................... 108

4-5 Data point positions in the 3x3 grid. Twelve 3x3 grids are measured long the
length of the cargo hold, every meter. .... ................................... 108

4-6 Tag readability test configuration. Tyvek sheet with 29 RFID tags (circled)
covering half of the cargo hold cross section.............................................. 109

4-7 Attenuation surface plots for each vertical slice of 3x3 data point at 433MHz,
circular antenna and top end antenna position............. ............................... 109

4-8 Attenuation surface plots for each vertical slice of 3x3 data point at 433MHz,
circular antenna and center ceiling antenna position................................... 110

4-9 Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz,
circular antenna and top end antenna position............. ............................... 110

4-10 Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz,
linear antenna and top end antenna position............................ ... ............... 110

4-11 Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz,
circular antenna and center ceiling antenna position................................... 111

4-12 Attenuation surface plots for each vertical slice of 3x3 data point at 2.45GHz,
circular antenna and top end antenna position ............... ..... ............. 111

4-13 Attenuation surface plots for each vertical slice of 3x3 data point at 2.45GHz,
linear antenna and top end antenna position ............... ............. ............... 111

4-14 Attenuation surface plots for each vertical slice of 3x3 data point at 2.45GHz,
circular antenna and center ceiling antenna position............................... 112

4-15 Distribution (in percentage) of each frequency tested, for circular antenna
only and two antenna locations. .............. ....... ............................ 113

4-16 Comparison of the change in average power levels and tag read rates for
both antennas through linear regression. .............. ....... ........................... 114

5-1 Diagram of the anechoic chamber setup for test 1. Four receiver antennas
shown for illustrative purposes. One receiver antenna is used at a time......... 130

5-2 Anechoic chamber set-up for test 2. A) The sample material is surrounded by
foam absorber and placed one wavelength from the emitting antenna. B) The
receiver antenna is taped behind the material sample. ............... .............. 130

6-1 ULD types and their respective tag positions. ........ ... ............ ................ 140









6-2 Temperatures recorded for top and bottom tags during flight (gate to gate).
Data is congregated by total flight time and type of aircraft............. ......... .... 141

6-3 Graph of averaged top and bottom tag temperatures during flight (gate to
gate) for both short flights (1-2h) to and from Montreal (YUL).............. ......... 142

6-4 Graph of averaged top and bottom tag temperatures during flight (gate to
gate) for both medium-short flights (4-6h) to and from Vancouver (YVR)....... 142

6-5 Graph of averaged top and bottom tag temperatures during flight (gate to
gate) for all 7-8h flights to and from London (LHR) or Frankfurt (FRA)........... 143

6-6 Graph of averaged top and bottom tag temperatures during flight (gate to
gate) for the longest flight (>9h) from London (LHR) to Vancouver (YVR)...... 143

6-7 Temperature profiles of all tags for ULDs AKH 1817 to and from Montreal...... 144

7-1 Overview of the air cargo operations where the circled steps represent
suggested RFID reading points. ............................. .......... .... .............. 159

A-1 Standard cargo compartment and containers for model DC-10 series 10,
10CF, 30, 30CF, 40 and 40CF. ........... ... ... ......................... ............... 162

A-2 Forward cargo loading door, model DC-10 series 10, 10CF, 30, 30CF, 40
a n d 4 0 C F ............. ......... .. .. ......... .. .. .. ...... ............................... 16 2

B-1 Attenuation surface plot example one slice of data (dBm)........................... 164

B-2 Attenuation surface plot for 433MHz, top end antenna position and circular
polarization. ............... .................. ........... ...... ......... ........... 165

B-3 Attenuation surface plot for 433MHz, center ceiling antenna position and
c ircu la r p o la riza tio n ........................................ ......................... 16 6

B-4 Attenuation surface plot for 915MHz, top end antenna position and circular
polarization. ............... .................. ........... ...... ......... ........... 167

B-5 Attenuation surface plot for 915MHz, top end antenna position and linear
polarization. ............... .................. ........... ...... ......... ........... 168

B-6 Attenuation surface plot for 915MHz, center ceiling antenna position and
c ircu la r p o la riza tio n ........................................ ......................... 16 9

B-7 Attenuation surface plot for 2.45GHz, top end antenna position and circular
polarization. ............... .................. ........... ...... ......... ........... 170

B-8 Attenuation surface plot for 2.45GHz, top end antenna position and linear
polarization. ............... .................. ........... ...... ......... ........... 171









B-9 Attenuation surface plot for 2.45GHz, center ceiling antenna position and
circular polarization ............................................... .......... 172

D-1 Signal strength surface plot example one slice of data (dBm).................... 179

D-2 Signal strength surface plot for 433MHz, top end antenna position and
circular polarization ............................................... .......... 180

D-3 Signal strength surface plot for 433MHz, center ceiling antenna position and
circular po larization ............................................... .......... 18 1

D-4 Signal strength surface plot for 915MHz, top end antenna position and
circular polarization ............................................... .......... 182

D-5 Signal strength surface plot for 915MHz, top end antenna position and linear
polarization. ............ ......... ... ............... ........................... ......... 183

D-6 Signal strength surface plot for 915MHz, center ceiling antenna position and
circular polarization ............................................... .......... 184

D-7 Signal strength surface plot for 2.45GHz, top end antenna position and
circular polarization ............................................... .......... 185

D-8 Signal strength surface plot for 2.45GHz, top end antenna position and linear
polarization. ............... .................. ........... ...... ......... ........... 186

D-9 Signal strength surface plot for 2.45GHz, center ceiling antenna position and
circular polarization ............................................... .......... 187

E-1 Temperature profiles of all tags for ULDs (A) AKH 2084 and (B) AKH 9778 to
and from Montreal (YUL) .................................... .............................. 188

E-2 Temperature profiles of all tags for ULDs AKH 1987 to and from Vancouver
(Y V R ) ........... ........... .. ........... .. ............................................... 1 8 9

E-3 Temperature profiles of all tags for ULDs (A) AKE 03782, (B) AKE 04090, (C)
AKE 04969, and (D) AKE 05335 to and from London (LHR)......................... 191

E-4 Temperature profiles of all tags for ULDs (A) AKE 03748, (B) AKE 04632, (C)
AKE 05168, and (D) AKE 05255 to and from Frankfurt (FRA) ...................... 193









LIST OF ABBREVIATIONS


A320 Airbus 320

A321 Airbus 321

A330 Airbus 330

AKE air cargo container prefix for LD3 without forklift holes

AKH air cargo container prefix for LD3-45

B777 Boeing 777

dB decibel

dBi decibel isotropic the forward gain of an antenna compared with
the hypothetical isotropic antenna, which uniformly distributes
energy in all directions

dBm decibel milliwatt power ratio in decibels (dB) of the measured
power referenced to one milliwatt (mW).

FAA Federal Aviation Administration

FCC Federal Communication Commission

FRA Frankfurt international airport

IC integrated circuit

LHR London Heathrow international airport

RBW resolution bandwidth

RF radio frequency

RFID radio frequency identification

SNR signal to noise ratio

ULD unit load device

YUL Montreal-Trudeau international airport

YVR Vancouver international airport

YYZ Toronto Pearson international airport









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

EXPLORATORY STUDY OF RFID APPLICATIONS FOR AIR CARGO OPERATIONS

By

Magalie Laniel

August 2010

Chair: Ray A. Bucklin
Major: Agricultural and Biological Engineering


The air cargo system is a complex network that handles a vast amount of freight

aboard passenger and all-cargo aircraft. With today's globalization, there is a growing

need for fresh products to be delivered year round all over the world, thus, temperature

sensitive items are likely to be shipped by air because of their relatively short shelf life.

Moreover, increasing demand for just in time delivery; containers being packed by third

parties; inspection time being very limited and transportation security being linked to

volume, monitoring of goods within containers as well as within the cargo warehouse

may contribute to enhanced security, operation efficiency and provide valuable real-time

information. New technologies to better track cargo shipments are accountable for

maintaining control and tracking along the supply chain. Radio frequency identification

(RFID) is seen as an emerging technology for improving the air cargo supply chain.

For RFID technology to be implemented, more research has to be done regarding

the environmental compatibility of air cargo warehouses, the regulations involved and

the materials encountered in this supply chain. Moreover, the frequency of choice may

be critical for system optimization. Therefore, the main objectives of this dissertation

are: identify the multiple RF interference encountered inside an air cargo warehouse;









evaluate the RF propagation behavior inside the cargo hold of an aircraft at different

frequencies; verify the effect of container wall materials on RF propagation; study the

temperature distribution of different cargo holds during flight.

The main findings of this research are that interference are lowest at 915MHz

inside the air cargo warehouses studied. Following the same direction, RF propagation

inside the cargo hold was found to be best at 915MHz when taking into account federal

spectrum regulations. In addition, the container materials experiment showed a very

strong effect of aluminum on RF transmission and minimal interaction for all other

composite materials. Moreover, there can be a significant temperature gradient between

the top and bottom of air cargo containers during ground operations as well as during

flights. The global system proposed from this research states that a combination of

active and passive tags at 915MHz could create a well suited structure for tracking of

the air cargo supply chain. To summarize, the findings of this dissertation suggest that

using 915MHz RFID systems for air cargo operations would lead to the most success

and system flexibility considering warehouse interference, cargo hold RF propagation,

temperature monitoring needs and types of tag technology available today.









CHAPTER 1
GENERAL INTRODUCTION

Automatic identification (Auto ID) of objects enables the organizations that

manage global supply chains to operate more efficiently and save cost. Auto ID includes

a host of technologies such as bar codes, smart cards, voice recognition, biometric

technologies and radio frequency identification (RFID). Bar codes have been the

primary means of identifying products since late 1960s. RFID offers many compelling

advantages over bar-codes, including non-line-of-sight operation, unique identifier,

higher read rate volumes and sensor capabilities, to name a few.

In addition, RFID technology enables computers to collect the unique ID assigned

to items. In combination with the Internet and associated infrastructure, RFID also

allows companies to track and trace individual items through the supply chain. RFID

aims to provide users a near-perfect supply chain visibility. That is, companies would be

able to know exactly where every item in their supply chain is at any moment in time.

The air cargo system consists of a large distribution network linking manufacturers

and shippers to freight forwarders to airport sorting and cargo handling facilities where

shipments are loaded and unloaded from aircraft (Figure 1-1). Business and consumer

demand for fast and efficient shipment of goods has fueled the rapid growth of the air

cargo industry over the past 25 years. World air cargo traffic is forecasted to expand at

an average annual rate of 5.8% for the next two decades, tripling current traffic levels

(Boeing, 2008). The air cargo supply chain has been looking at RFID as a solution to

increase its safety, operation efficiency and monitoring capability for many years (Figure

1-2). Today's mostly manual processes (accepting, weighing, dimensioning, sorting,

storing, building and breaking down shipments, etc.) are not keeping up with the









growing demand for fast and reliable shipping services. In addition, no time stamp is

provided each time a shipment is loaded or unloaded from an aircraft, and only manual

inspection tells if the shipment is in the right aircraft or not. In air cargo operations,

shipments are still being lost and items sometimes travel without their associated

documents, which leads to claims that the carrier has to pay. Moreover, real time

locating of loose goods as well as unit load devices (ULD) in and out of the air cargo

warehouse can provide visibility that not only the shipping company could benefit from;

but also customers see value in knowing where their shipments are.

The tracking and rapid locating of baggage, loose freight and containers

(especially the associated integrity assurance of those items) is also essential to the

overall security of a commercial flight. This tracking / locating of goods is accomplished

today, for the most part, only by a very labor-intensive manual process. The introduction

of RFID to provide this asset tracking and locating offers the opportunity for: centralized

monitoring; continuous surveying; automatic event logging; and, of course, more rapid

finding of items when retrieval is mandatory (Cerino and Walsh, 2000).

When time becomes a primary consideration for delivery, air transportation is the

mode of choice. According to McCarthy (2003), one of the key drivers for the use of air

cargo over other modes is the weight to value ratio of shipped goods. Some specific

market segments include: extremely high value products such as jewelry, luxury

automobiles, and race horses; just-in-time products such as electronics and auto parts;

perishables such as fresh foods, flowers, and seasonal apparel; and time-sensitive

products such as medical supplies and pharmaceuticals. Many pharmaceutical and

biotech products have a correlated sensitivity to temperature and high value (Wright,









2008) which makes these industries a major customer of the air cargo industry.

Moreover, the quality and integrity of pharmaceutical products can be vital to people's

lives; therefore, temperature management of such shipment is of prime importance.

With today's globalization, there is a growing need for fresh products to be

delivered year round all over the world and with the cold supply chain requiring fast

delivery; more and more perishable items are being shipped by air (Vega, 2008).

Unfortunately, a faster transit time does not always imply controlled temperature

throughout transportation. Of approximately 2.6 million tons of perishables air freighted

in 2008, nearly 30% was estimated to be lost due to handling and temperature abuse

(Catto-Smith, 2006).

RFID technology can also be combined with many different sensor applications,

such as monitoring temperature, humidity, motion, etc. These features, with real-time

tracking of unique IDs throughout the air cargo supply chain open numerous valuable

opportunities for shippers and customers. In essence, RFID is revolutionizing the way

products and goods are tracked and traced in the supply chain.

It has been shown that RFID can significantly improve warehouse operation

efficiency and supply chain performance (Chow et al., 2006; Poon et al., 2009;

Veronneau and Roy, 2009; Visich et al., 2009; Wang et al., 2010). It has also been

shown that RFID can improve the overall quality and shelf life of perishables through the

cold supply chain (Emond, 2007; Jedermann et al., 2007, 2009; Ruiz-Garcia et al.,

2008; Abad et al., 2009). Although, to implement such technology in the air cargo

industry, more research has to be done regarding many compatibility aspects of RFID

technology and the air cargo world.









Commercial RFID systems are available under different standards, which work at

different frequencies. Choosing the best suitable frequency for an application depends

on many factors. For instance, the read range needed, the size of tags preferred and

the type of environment surrounding the RFID system (materials and other

interferences. In this research, three frequencies will be evaluated: 433MHz, 915MHz

and 2.45GHz. Those frequencies are thought to be the most appropriate for the air

cargo world today, mostly because of their longer read ranges.

Objectives. The main goal of this work is to evaluate the possibilities of using

RFID to improve air cargo operations in general, as well as for perishable

transportation. More specifically, some questions associated with air cargo operations

will be addressed (Figure 1-3). Therefore, the four main objectives of this dissertation

are:

* Measure and evaluate the interference level at 433MHz, 915MHz and 2.45GHz in
air cargo warehouses (chapter 3).

* Obtain 3D mapping of RF propagation inside a cargo hold at 433MHz, 915MHz
and 2.45GHz and compare with RFID tag readability at 915MHz (chapter 4).

* Evaluate RFID behavior around five air cargo container (ULD) materials at
433MHz, 915MHz and 2.45GHz (chapter 5).

* Study the temperature distribution inside air cargo containers in different cargo
holds and aircraft during flight (chapter 6).



























Figure 1-1. Brief overview of the air cargo supply chain.


Warehouse

Problematic:
*Item locating
*Pieces/ULD
association
*Operation
efficiency
(manual)
*No time stamp
(for goods
movement)
*No monitoring
(perishable)


Aircraft


Shipments


Problematic:
*Security (during
flight)
*No time stamp of
load/unload of
ULD
*No monitoring
(perishable)


Airport


Figure 1-2. Some problems associated to the air cargo warehouse and aircraft
operations.


P


11 -ego












Warehouse


Questions:
*RF environment
compatibility?
*Frequency
choice? Shipments
*Container
materials RF
lucent or not? \


Questions:
*RF propagation
vs. frequency?
*Air cargo
material RF
friendliness?
*Temperature
distribution, is an
issue?


Airport


Figure 1-3. Some questions associated with warehouse and aircraft air cargo
operations.


Aircraft









CHAPTER 2
RFID IN AIR CARGO: A LITERATURE REVIEW

Introduction

This review will give a general idea of radio frequency identification (RFID)

technology, its definitions and the way systems work. Then will follow an overview of the

air cargo world, aviation history, the current market status, plus a brief description of the

major parts and their operations. In addition, aircraft construction, systems, avionics and

safety issues will be discussed. Subsequently, the subject of RFID in aviation will be

described through regulations, applications as well as possible interference of the

technology. The concluding remarks comment on the potential of RFID to improve air

cargo operations in general.

RFID Technology and Definitions

Radio frequency identification (RFID) is an automatic wireless data collection

technology with a long history. The fundamentals of RFID technology are based on

electromagnetic energy studies, originating with Michael Faraday's explanation of light

and radio waves as forms of electromagnetic energy back in 1846. For the last two

decades RFID tags have been used in many applications (e.g. automatic toll roads,

smart cards, store theft protection, access control, animal tracking, item tracking, etc.)

which supply chain management and item tracking have been the fastest growing areas

(Landt, 2005). Improvements in semiconductor technology resulted in reduction in the

size of circuitry, reduction in cost of tags, increased functionality, and increased

reliability, which sped up the industrial applications of RFID (Landt, 2005).









Radio Waves

Radio waves account for a portion of the electromagnetic spectrum (Figure 2-1).

Radio waves at their most basic are considered as wave forms of electrical and

magnetic fields and as a result, have amplitude, wavelength (X), velocity (v) and

frequency (f), the relationship of which is expressed as:

v = 3 f (2-1)


Electromagnetic waves are created by electrons in motion and consist of oscillating

electric and magnetic fields. These waves can pass through a number of different

material types (Lahiri, 2006).

The highest point of a wave is called a crest, and the lowest point is called a

trough, as shown in Figure 2-2. The distance between two consecutive crests or two

consecutive troughs is called the wavelength. One complete wavelength of oscillation of

a wave is called a cycle. The time taken by a wave to complete one cycle is called its

period of oscillation. The number of cycles in a second is called the frequency of the

wave. The frequency of a wave is measured in hertz (abbreviated as Hz) and named in

honor of the German physicist Heinrich Rudolf Hertz. If the frequency of a wave is 1 Hz,

it means that the wave is oscillating at the rate of one cycle per second. It is common to

express frequency in KHz (or kilohertz = 1,000Hz), MHz (or megahertz = 1,000,000Hz),

or GHz (or gigahertz = 1,000,000,000Hz). Amplitude is the height of a crest or the depth

of a trough from the undisturbed position (Lahiri, 2006).

Radio waves can be further divided up into groups; Low Frequency (LF), High

Frequency (HF), Ultra High Frequency (UHF) and Microwave Frequency (MF) with

similar categories applying to RFID systems. Electromagnetic energy has been best









described as a stream of photons each traveling at the speed of light in a wave like

pattern. An electromagnetic wave propagates in a direction that is at right angles to the

vibrations of both the electrical and magnetic oscillating fields (Winder and Carr, 2002).

Radio waves and microwaves are situated towards the lower end of the electromagnetic

spectrum (Figure 2-1) meaning that waves situated in the low frequency category

posses lower amounts of energy (E) compared to microwave frequency waves

according to Planck's equation:

E= h f (2-2)


Where h is Planck's constant.

Electromagnetic waves can be characterized in terms of frequency, wavelength or

energy (as shown using Planck's relation above). Taking the speed of light (c = 3x108)

as the velocity, it is now possible to say (Meyers et al., 2007):

c = .f, A=7 or f= (2-3)
I A


Polarization of electromagnetic waves

The polarization of an electromagnetic wave is determined by the direction of the

electric field of the wave. There is a difference between linear polarization and circular

polarization. In linear polarization the direction of the field lines of the electric field in

relation to the surface of the earth provide the distinction between horizontal (the

electric field lines running parallel to the surface of the earth) and vertical (the electric

field lines running at right angles to the surface of the earth) polarization (Finkenzeller,

2003).









The transmission of energy between two linear polarized antennas is optimal if the

two antennas have the same polarization direction. Energy transmission is at its lowest

point, on the other hand, when the polarization directions of transmission and receiving

antennas are arranged at exactly 90 or 2700 in relation to one another (e.g. a horizontal

antenna and a vertical antenna). On the other hand, circular polarization occurs when

the polarization direction of the electromagnetic field generated rotates through 3600

every time the wave front moves forward by a wavelength. The rotation direction of the

field can be determined by the arrangement of the delay line. We differentiate between

left-handed and right-handed circular polarization (Finkenzeller, 2003).

Electromagnetic waves properties

In free space, all electromagnetic waves obey the inverse-square law which states

that the power density of an electromagnetic wave is proportional to the inverse of the

square of the distance from the source. In other words, as the separation distance is

doubled, the electrostatic force is decreased by a factor of four (Henderson, 2010). As

the electromagnetic waves propagate in their environment, they encounter many

objects and behave differently around those obstacles, sometimes in a critical way

towards the communication link needed for the functioning of an RFID system. It has

been reported that environmental factors may decrease the reader range of passive

RFID systems by at least 50% (Keskilammi et al., 2003). It is also well established that

higher frequencies experience greater attenuation levels than lower frequencies

(Keskilammi et al., 2003). Any wave incident upon an object will penetrate the material,

a portion may be transmitted and another portion may also be reflected back into the

environment. The exact amount of transmission and reflection is also dependant on the

angle of incidence, material thickness, and dielectric properties (Blaunstein and









Christodoulou, 2007). Part of the high frequency energy that reaches the object is

absorbed by the object and converted into heat; the rest is scattered in many directions

with varying intensity (Finkenzeller, 2003).

Radio waves can be affected by the material through which they propagate. A

material is called RF-lucent for a certain frequency if it lets radio waves at this frequency

pass through it without any substantial loss of energy. A material is called RF-opaque if

it blocks, reflects, and scatters RF waves. A material can allow the radio waves to

propagate through it but with substantial loss of energy. These types of materials are

referred to as RF-absorbent. The RF-absorbent or RF-opaque property of a material is

relative, because it depends on the frequency. That is, a material that is RF-opaque at a

certain frequency could be RF-lucent at a different frequency (Lahiri, 2006).

The presence of more than one wave in a space may result in interference

between the waves, which can be constructive (they reinforce one another), or

destructive (cancel each other in whole or in part). There are a number of different ways

an electromagnetic wave may interact with materials in its surrounding area as follows

(Wu et al., 2006; Domdouzis et al., 2007):

Scattering: This occurs when a wave hits an obstacle smaller than its wavelength.

It leads to the formation of scattered waves which are redirected with random phase

and amplitude (Blaunstein and Christodoulou, 2007). This can be the result of rough

surfaces, small objects or irregularities in the transmission medium.

Refraction: This is the change in direction of a wave due to a change in its speed.

This is most commonly observed when a wave passes from one medium (with a certain









refraction index) to another at an angle. Refraction is described by Snell's law, which

states that the angle of incidence is related to the angle of refraction (Reed, 2009).

Fading: This is a variation of the signal strength with time (Meyers et al., 2007). It

occurs due to time dependent changes in multipath. Fade zones are small areas inside

the interrogation zone that lead to periodic attenuation of the received signal. This effect

increases with the distance from the emitting antenna. This occurrence is too random to

make possible the prediction of signal strength at a particular point in time (Mac Carthy,

2009).

Multipath: This occurs when a radio wave arrives at a particular receiving antenna

from more than one propagation route due to its interactions with the surrounding

environment (Lahiri, 2006). Multipath, or path loss strongly depends on propagation

environment.

Reflection and cancellation: The electromagnetic field emitted by the reader is

not only reflected by a transponder, but also by all objects in the vicinity, the spatial

dimensions of which are greater than the wavelength of the field (Rappaport, 2002).

The reflected fields are superimposed upon the primary field emitted by the reader. This

leads alternately to a local damping or even so-called cancellation (anti-phase

superposition) and amplification (in-phase superposition) of the field at intervals of A/2

between the individual minima. The simultaneous occurrence of many individual

reflections of varying intensity at different distances from the reader leads to a very

erratic path of field strength around the reader, with many local zones of cancellation of

the field. Such effects should be expected particularly in an environment containing

large metal objects (Finkenzeller, 2003). The importance of these properties cannot be









over emphasized as they are all hugely important in relation to passive UHF RFID

systems as they impact on how the electromagnetic wave (essential for coupling) is

affected by different objects.

Direct penetration: Little or no reflection occurs when electromagnetic waves

penetrate directly through objects such as paper, non conductive plastics or textiles

(Penttila et al., 2006). These materials, including most composites, are non-absorbing

and possess low refractive indexes. Such materials are generally referred to as being

RF-lucent.

Diffraction: Similarly to light propagation, materials surrounding radio waves can

provoke diffraction of the waves, which can lead to RF signal variations. It is described

as the apparent bending of waves around small obstacles, the spreading out of waves

past small openings or the deviation in the path of a wave that encounters the edge of

an obstacle. At high frequencies, diffraction, like reflection, depends on the geometry of

the object, as well as the amplitude, phase, and polarization of the incident wave at the

point of diffraction (Rappaport, 2002)

RFID System Overview

A basic RFID system consists of a computer with software connected to a reader

and one or more reader antennas, which communicate wirelessly with tags (Figure 2-3).

The reader transmits an RF signal to the tags via its antenna(s). The tags receive power

from the antenna and then send their information back. Following is a description of

each component in more detail.

Readers

RFID readers can be portable or fixed, depending on the application. A typical

RFID reader has both transmitting and receiving functions for data transfer and









communication with tags (Keskilammi et al., 2003; Poussos and Kostakos, 2009). Most

interrogators consist of an RF transceiver module (transmit and receive), a signal

processor, a controlling unit and a coupling element (antenna) and data interface to a

host system (Lahiri, 2006).

EIRP: The Equivalent Isotropic Radiated Power (EIRP) determines the power of

the signal transmitted by the reader in the direction of the tag. Maximum allowed EIRP

is limited by national regulations (e.g. in North America it is 4W) (Nikitin and Rao, 2006).

Reader sensitivity: is another important parameter which defines the minimum

level of the tag signal which the reader can detect and resolve. The sensitivity is usually

defined with respect to a certain signal-to-noise ratio or error probability at the receiver.

Factors which can affect reader sensitivity include receiver implementation details,

communication protocol specifics, and interference, including signals from other readers

and tags. An ideal reader can always detect an RFID tag as long as the tag receives

enough power to turn on and backscatter (Nikitin and Rao, 2006).

Antenna

Antennas are essential components of both RFID tags and readers. Their principal

function is the facilitation of a dual directional communication link between tag and

reader (Dobkin, 2008). At its most basic an antenna is a particular arrangement of

conductors designed to transmit an electromagnetic field in response to the application

of an alternating electric current. It also has the ability to generate a voltage between

terminals when placed in a time varying electromagnetic field (Finkenzeller, 2003). Both

RFID tags and reader antennas come in a variety of sizes and shapes which

determines their operational characteristics (polarization for example).









Reader antenna: These are used to communicate with the nearby tags. Antennas

have emitting and receiving capabilities. The antenna first propagates the RF wave into

the environment in order to establish a communication link between the tags and reader

to facilitate coupling (Liang et al., 2006). This RF wave creates an interrogation zone

which is an area surrounding the antenna where communication will take place provided

that a tag is present. Once communication is established, the antenna also receives a

resultant signal from the tag which is transferred to the reader for demodulation.

The antenna may be physically incorporated into the reader, which is generally the

case for handheld units. Alternatively, the antenna might be individually housed and

attached to the reader by appropriate cables (Roussos and Kostakos, 2009). Depending

on the desired use, readers can support connection to more than one antenna at a time.

Tag antenna: They operate on the same principle as reader antennas, but face

some different practical challenges (Dobkin, 2008). In the case of passive RFID tags,

the tag antenna is responsible for receiving the electromagnetic wave from the reader

antenna and reflecting the modulated backscatter signal to the reader. For active RFID

tags, the antenna is responsible for emitting the internally generated signal. The size

and shape of the antenna determines the operating frequency as well as the application

and ideal orientation of the tag (Fuschini et al., 2008). Tag antenna must also be small

enough to fit the item it is identifying; have omnidirectional and hemispherical coverage;

have a polarization that matches with the reader; and be robust and low cost

(Keskilammi et al., 2003).

Polarization: Propagating electric fields point in a certain direction in space.

Polarization refers to the orientation of the electric field radiated by the antenna. If the









vector rotates with time, then the wave is elliptically polarized. The degree of ellepticity

from a circle to a straight line gives circular and linear polarization. For linear

polarization (vertical / horizontal), the vector oscillate on one plane as the wave

propagates whereas in circular polarization it rotates through 360 degrees per cycle

(Figure 2-4).

Gain: The gain, or amplification factor, is the factor by which the input power is

amplified. Ideally a reader antenna has a high gain due to the fact that the received

power at the tag is directly proportional to the reader antenna gain.

Tags

An RFID tag is a device that can store and transmit data to a reader in a

contactless manner using radio waves. Tags have three main components: an

integrated circuit (IC or chip), an antenna and a substrate (Meyers et al., 2007). They

are available in a wide variety of sizes, shapes and functionality which determines their

unit cost (Jedermann et al., 2009). The tag is responsible for storing and sharing user

defined information regarding the item to which it is attached. It may be constantly in an

active state whereby it may record storage conditions in its immediate environment or it

may remain in a dormant state until activated by an interrogating wave from a nearby

reader depending on its classification (Nikitin and Rao, 2006). It communicates with the

reader by superimposing its stored data (through signal coding and demodulation).

Transponders can be classified according to sources of energy:

* Passive tags
* Active tags
* Semi-active tags
* Semi-passive tags









Passive tags: Passive RFID tags have no internal power source (no battery). In

inductively coupled systems, when the tags are present in the RF field of an RFID

interrogator, the energy induced on the tag circuitry is used for transmitting back the ID

of the tag. In UHF systems, electromagnetic backscatter coupling is used at the tag

circuitry for changing the impedance of the tag antenna according to its ID. A passive

tag is simple in its construction and has no moving parts. As a result, such a tag has a

long life and is generally resistant to harsh environmental conditions. Passive tags can

be very small and low cost to manufacture. On the other hand, they have limited data

capacity and shorter read range. In tag-to-reader communication for this type of tag, a

reader always communicates first, followed by the tag. The presence of a reader is

mandatory for such a tag to transmit its data (Lahiri, 2006).

Active tags: Active tags are beaconing in a defined period of time by using

integrated power supplies (battery for example). It does not need the reader's emitted

power for data transmission. The on-board electronics can contain microprocessors,

sensors (temperature, humidity, motion, etc.), and input / output ports powered by the

on-board power source. Therefore, for example, these components can measure the

surrounding temperature and generate the average temperature data (Keskilammi et

al., 2003). The components can then use this data to determine other parameters such

as the expiry date of the attached item. The tag can then transmit this information to a

reader. In tag-to-reader communication for this type of tag, a tag always communicates

first, followed by the reader (Lahiri, 2006). Active tags also have longer read ranges

than passive tags. They are ideal in environments of high electromagnetic interference

because of their ability to broadcast a stronger signal with the aid of their internal power









source (Jeddermann et al., 2009). Some advantages and disadvantages of active tags

are:

* Increased functionality (sensor, monitoring, recording)
* Long read ranges
* Large memory capacity
* Physically bulky
* High production costs
* Fragile because of moving parts in the design
* Life span limited to power source

These tags are commonly used in RTLS (Real Time Location Systems) in which the tag

continuously reports its ID to the receiver units and location of the tag is usually

calculated by using RSS (Received Signal Strength) information and triangulating

between different receivers (Altunbas, 2010).

Semi-active tags: The name associated with this type of tag is not yet widely

accepted. It is somewhat a type of active tag that enters a sleep or a low-power state in

the absence of interrogation by a reader. A reader "wakes up" such a tag from its sleep

state by issuing an appropriate command. This state saves the battery power, and

therefore, a tag of this type generally has a longer life compared to an active transmitter

tag. In addition, because the tag transmits only when interrogated, the amount of

induced RF noise in its environment is reduced.

Semi-passive tags: These combine both passive and active type technology.

Semi-passive systems have internal batteries but are not beaconing signals in a defined

period. In the presence of the RF field of an interrogator, the tag wakes up and starts to

backscatter its ID to the interrogator (like a passive tag); but using its integrated battery

supply to increase the signal (McCarty, 2009). The trade off for energy efficiency is to









have reduced response time caused by the time slot needed to wake up the

transponder.

Tag generation and classification. Tags can be classified according to their

power source, as seen earlier, but they may also be grouped according to their

functionality (Meyers et al., 2007). This classification is based on EPCglobal standard

as shown in Table 2-1. This classification of tag is based on the format, read / write

capability and programming capability. The EPC classification consists of Class and

Generation. The Class describes a tags basic functionality, for example whether it has

memory or an on-board power source, whereas Generation refers to a tag

specification's major release or version number (Khan et al., 2009).

Frequencies

Commercial RFID systems designed for different applications can work at different

frequencies, such as:

Low frequency (LF): Frequencies between 30kHz and 300kHz are considered

low, and RFID systems commonly use the 125kHz to 134kHz frequency range. At LF,

the power supply to the transponder is generated by inductive coupling, which means

short read ranges between the tags and reader antenna. RFID systems operating at LF

generally have low data-transfer rates from the tag to the reader, and are especially

good if the operating environment contains metals or liquids (Lahiri, 2006), which is why

LF systems are commonly used in animal identification applications.

High frequency (HF): HF ranges from 3MHz to 30MHz, with 13.56MHz being the

typical frequency used for HF RFID systems. A typical HF RFID system uses passive

tags, has a slow data-transfer rate from the tag to the reader, and offers fair

performance in the presence of metals and liquids (Lahiri, 2006). Moreover, HF systems









also work within short read ranges (generally within 1m), it is used in smart cards

applications like access control or contactless payments.

Ultra-high frequency (UHF): UHF is the most common passive RFID tags used

in supply chain applications worldwide. While the entire UHF spectrum ranges from

300MHz to 1GHz, it operates between 902-928MHz in the Americas; 865-868MHz in

Europe, middle east and Russia; and 866-869MHz and 923-925MHz in Asia, Australia

and the Pacific. The accepted standard for passive UHF frequency is ISO 18000-6C

(UHF Gen2). Active or semi-active RFID systems in UHF frequencies operate at

433MHz. ISO 18000-7 is the accepted standard for parameters of active air interface

communications at 433MHz. A UHF system can therefore use both active and passive

tags and has a fast data-transfer rate between the tag and the reader, but performs

poorly in the presence of metals and liquids (not true, however, in the cases of low UHF

frequencies such as 433MHz) (Lahiri, 2006). UHF systems working in the

electromagnetic field offer a much longer read range than lower frequencies.

Microwave frequency: Microwave frequencies range upward from 1GHz. A

typical microwave operating frequency is either at 2.45GHz or 5.8GHz in the Industrial

Scientific and Medical (ISM) band. Microwave systems can be either passive or semi-

passive and provide the fastest data communication rates compared to the other

frequencies (Lahiri, 2006). Microwave frequency, like UHF, offers long read ranges,

especially when working in an active RFID system. Read range performance for passive

systems in the presence of water and metallic surfaces is very poor because of the

higher signal attenuation at higher frequency (Friis, 1946).









National licensing regulations. In the USA, RFID systems must be licensed in

accordance with licensing regulation FCC Part 15 from the Federal Communications

Commission. This regulation covers the frequency range from 9kHz to above 64GHz

and deals with the intentional generation of electromagnetic fields by low and minimum

power transmitters (intentional radiators) plus the unintentional generation of

electromagnetic fields (spurious radiation) by electronic devices such as radio and

television receivers or computer systems. The category of low power transmitters

covers a wide variety of applications, for example cordless telephones, biometry and

telemetry transmitters, on-campus radio stations, toy remote controls and door openers

for cars. Inductively coupled or backscatter RFID systems are not explicitly mentioned in

the FCC regulation, but they automatically fall under its scope due to their transmission

frequencies, which are typically in the ISM bands, and their low transmission power

(Finkenzeller, 2003). Table 2-2 lists some of the frequency ranges that are important for

RFID systems.

Air Cargo

Aviation History

Times have changed since the Wright Brothers and the first flight of a powered

aircraft in 1903 (Taylor 1989; Bilstein 1994; Wegener 1997). They were not only

pioneers of flight, they also were the first ones to ship goods by air, when in November

1910, a department store from Ohio made arrangement with them to have a bolt of silk

flown up from Dayton to Columbus (Bilstein, 1994). Not long after, in 1911, the first

official airmail flight is made in India, where 6,500 letters were carried over about 10 km

(Taylor, 1989). It is only after World War I, on May 15, 1918 that airmail service began

in the US between New York City, Philadelphia and Washington D.C. with a JN-4, which









was built as an Army training airplane. In 1919, American Railway Express made an

unsuccessful attempt to deliver cargo to Chicago (Bilstein, 1994). In the mean time, the

Airmail Act (1925) and the Air Commerce Act (1926) were created. Later, in 1927, plans

were made for four commercial airlines (National Air Transport, Colonial Air transport,

Boeing Air Transport and Western Air Express) to fly express. From hesitant beginnings

to slow progress, commercial air cargo made great strides in the post-World War II era.

Wartime experience in long-range cargo operations helped, but the postwar availability

of dozens of surplus military multiengine transports was more important. Unfortunately,

the maintenance cost of those aircraft only allowed a handful of companies to survive.

On the other hand, the scheduled passenger lines, sensing lost revenues, began to pay

more attention to cargo services in their normal passenger routes and began to operate

their own all-cargo services (Bilstein, 1994).

Air Cargo Supply Chain

The air cargo system consists of a large, complex distribution network linking

manufacturers and shippers to freight forwarders to airport sorting and cargo handling

facilities where shipments are loaded and unloaded from aircraft (Elias, 2007).

The airport forms an essential part of the air cargo supply chain, because it is the

physical site at which a modal transfer of transport is made from the air mode to land

mode. It is the point of interaction between the airline and the user (Ashford et al.,

1983). Airports are divided into landside (parking lot, access roads, etc.) and airside (all

areas accessible to aircraft, including runways, taxiways and ramps) areas. In addition

to people, airports move cargo around the clock. Cargo airlines often have their own on-

site and adjacent infrastructure to transfer parcels between ground and air.









Air cargo warehouse operations

Figure 2-5 presents the layout of an airline's new cargo facility. The cargo terminal

is divided into an import area and an export area. The flow of goods through the

terminal is either from the airside to the landside (terminating freights or connecting

freights requiring the road feed service), from the landside to the airside (originating

freights or connecting freights arriving from a road feeder service), or from the airside to

the airside via the terminal (connecting freights).

Export area: The export area is dedicated to receiving, processing and preparing

outbound freights, which refers to all shipment moving from an outside customer, and

going onto a flight. All freight arrives at the cargo facility from the landsidee export"

area, either as bulk or as shipper loaded unit device. The freight gets weighed and

dimensioned by the acceptance agent and stored at the appropriate location depending

on its flying time and destination. If items are bulk, they ultimately go to the build-up

area to be put in a ULD (Unit Load Device) or are transported in a tub cart directly to the

airplane if this airplane is bulk loaded. ULDs are transported onto roller system through

the cargo facility and onto trailers to the airside. All export shipments leave the

warehouse via the "airside export" doors.

Import area: The import area is dedicated to receiving, processing and releasing

inbound freights which refers to all shipments coming from a flight, going to an outside

customer. ULDs are transported the same way between the airside and cargo facility

(trailers). Bulk is unloaded from the aircraft directly into tub carts. Everything is brought

back to the cargo warehouse via the "airside import" area, is broken-down when needed

and stored until customer pick-up.









The movement of transiting goods (from one flight to another flight) also goes

through the warehouse. It is considered "import" as it enters via the import airside, and

becomes "export" as is it processed in the cargo facility and moves to the export side

before exiting the warehouse through the "airside export" doors to reach its next flight.

Unit load device (ULD)

Unit Load Devices (ULDs) play a vital part in ensuring that as air-cargo volumes

increase; they are moved safely, quickly and cost-effectively (IATA, 2002). ULD is the

correct terminology used by the air transport industry for containers and loading units

that are used for the carriage of cargo by air. It allows large quantities of cargo to be

bundled into large units. Pallets and nettings as well as rigid containers are commonly

used for freight transport by air. Each ULD is required to have a marking that identify its

type code, maximum gross weight and actual tare weight (IATA, 2002). Currently,

technical specifications for unit load devices are set by the International Air Transport

Association (IATA).

While the world is talking about climate change, the airline industry is looking at

ways to be more fuel efficient to minimize their operational costs as well as their impact

on the environment. One way to do so is to reduce weight, minimizing weight without

compromising the business volume is feasible by using lighter containers, or ULDs.

Composite ULDs can save up to 25% of the tare weight of a traditional aluminum ULD

(Nordisk, 2010). Kevlar ULDs are constantly replacing older aluminum containers and

account for approximately 39% of a major airline's ULD fleet. Aluminum ULDs still add

up to 43% of their fleet, whereas Lexan containers count for the remaining 18%.









Market

During the late 1960s, the total tone kilometerage of freight doubled every four

years, an average annual growth rate of 17%. At that time, the aviation world was

replete with extremely optimistic forecasts of a burgeoning air cargo market. The

prolonged and recurrent economic recessions and the tenfold increase in oil prices of

the 1970s militated against sustained growth in North America (Ashford et al., 1983).

Today's trend. World air cargo traffic grew 5.1% in 2007, which followed 3.2%

growth in 2006 and 1.7% growth in 2005, making those three years the weakest growth

period for the industry since the first Gulf War, 1990-1992. Tepid traffic growth can be

largely attributed to high fuel prices, which were increasing from late 2003 through July

2008. In response to the ongoing rise in jet fuel prices, freighter operators have

accelerated fleet renewal activities, most notably in the large wide-body sector.

Freighters count for about 10% of the total airplane fleet (Boeing, 2008). A wide-body

aircraft is a large airliner with two passenger aisles, whereas a narrow-body only has

one passenger aisle.

A few decades ago, it was hard to foresee the present degree of traffic volume in

aviation. An increase of up to 5.8% per year is estimated for the next two decades

which will mean triple the amount of the present cargo traffic volume in the next 20

years (Boeing, 2008). As a result of this increase, the focal point in aviation research

has changed: Socio-economic aspects are coming to the fore. The reduction of

emissions such as noise and pollutants is becoming more significant. In particular, the

reduction of the weight of the structure of future aircraft is a central task that will enable

a reduction of fuel consumption and an increase in the payload (Wilmes et al., 2002).









The international competitive situation and the related increase in global

competition in the aircraft industry is additionally making it necessary to considerably

reduce costs in the development, production, and maintenance of the next generation of

aircraft. The development time must be considerably shortened in order to enter aircraft

faster into service. In addition to weight and cost, additional challenges in the future will

be increased safety requirements for aircraft in the case of accidents, etc.

Improvements in these areas are indispensable in order to ensure a high acceptance of

this means of transportation in the future (Wilmes et al., 2002).

Materials in Commercial Aircrafts

Commercial aircraft include types designed for scheduled and charter airline

flights, carrying both passengers and cargo. The larger passenger-carrying types are

often referred to as airliners, the largest of which are wide-body aircraft. Some of the

smaller types are also used in general aviation. Aircraft construction materials used to

be mostly aluminum alloys, but nowadays more and more composites are utilized in

aircraft design.

Composites

The development of composite materials is considered to be one of the most

important advances in aviation design since aluminum was introduced in the 1920s.

Development of various composite materials has had a very positive impact on the

performance, shape, reliability, weight, cost and composition of modern aircraft.

Composites are a combination of two or more significantly different inorganic or organic

components. Although the components together form a composite material they each

maintain their original form and do not blend together. In a composite material, one

component serves as a "matrix", being the component that holds everything together,









with the other component or components serving as reinforcement. An epoxy resin

matrix with glass fiber reinforcing is one of the more commonly known composite

materials, but continuing research is resulting in the production of various other

composite materials which are proving beneficial in aviation design as well as in other

industries (Anonym, 2007). Despite their strength, light weight, long life expectancy,

corrosion resistance, and resistance to damage from cyclic loading (fatigue);

composites have not been a miracle solution for aircraft structures. Composites are

hard to inspect for flaws and can sometimes be brittle. Some of them absorb moisture.

Most importantly, they can be expensive, primarily because they are labor intensive and

often require complex and expensive fabrication machines. Aluminum, by contrast, is

easy to manufacture and repair (Day, 2009).

Modern airliners use significant amounts of composites to achieve lighter weight.

About 10% of the structural weight of the Boeing 777, for instance, is composite

material (Day, 2009). The new Boeing 787 Dreamliner has made extensive use of

composite materials, resulting in a lighter weight airplane which is expected to have a

number of benefits including greater fuel efficiency. This twin-engine, wide-body jet

airliner is constructed from 50% composite with aluminum, titanium and steel making up

45% and a variety of components making up the balance of 5%. The composite material

most used in the Boeing 787 Dreamliner's construction is carbon fiber reinforced plastic

(Boeing, 2010a).

Metals

Aluminum still remains a remarkably useful material for aircraft structures and

metallurgists have worked hard to develop better aluminum alloys (a mixture of

aluminum and other materials). Aluminum is a very tolerant material and can take a









great deal of punishment before it fails. It can be dented or punctured and still hold

together (Day, 2009). Aluminum alloys used in the aerospace industry are high strength

and able to perform well in harsh and challenging environments. 7075 Aluminum is the

alloy of choice when it comes to manufacturing aircraft parts, and 5052 aluminum,

which is not quite as strong but has more weldability, is sometimes used. 7075 contains

zinc and copper, which is ideal for highly stressed parts and is considered the strongest

type of aluminum. It has good high temperature resistance and corrosion resistance,

both necessary characteristics in aircraft aluminum. Aircraft metal must be strong yet

lightweight at the same time, and aluminum exhibits a good strength-to-weight ratio,

making it the first choice in airplane construction. The airframe of a typical commercial

transport aircraft is 80% aluminum by weight (The Aluminum Association, 2008).

Electrical Systems in Commercial Aircrafts

Aircraft power can be generated by DC or AC power sources. DC systems usually

supply 28 VDC at all times. Most of today's commercial aircraft use AC power sources,

which are three-phase systems where three sine waves are generated 120 degrees out

of phase from each other. In this layout the phase voltage of a standard aircraft system

is 115 VAC and the standard frequency is 400Hz; which is the same standard for

ground power at most airports (Moir and Seabridge, 2001).

Avionics: This is a portmanteau word of "aviation electronics". It comprises

electronic systems for use on aircraft, comprising communications, navigation and the

display and management of multiple systems. Table 2-3 lists the main systems and their

respective working frequencies.









Temperature Profile in Commercial Aircraft

Temperature is well regulated in the cabin of most passenger flights, but it is not

necessarily the case inside the cargo hold or of freighter flights. Temperature

distribution around cargo depends on many factors, to name a few: weather, duration of

flight, type of aircraft (ability to control cargo ambient temperature), altitude, transit time

(on tarmac), etc. As shown in Figure 2-6, a study on an international shipment of live

mice during summer showed that both heating problems (during airport handling) and

cooling problems (during flight) can occur (Syversen et al., 2008).

Aircraft Safety

Every day approximately six million people board airplanes and arrive safely at

their destinations. Flying is one of the safest modes of transportation today. The overall

safety record of commercial airplanes is excellent and has been steadily improving over

time. During the 1950s and 1960s, fatal accidents occurred about once every 200,000

flights. Today, the worldwide safety record is more than ten times better, with fatal

accidents occurring less than once every 2 million flights (Boeing, 2010b).

Cargo security and monitoring

The air cargo system is a complex, multi-faceted network that handles a vast

amount of freight, packages, and mail carried aboard passenger and all-cargo aircraft.

The air cargo system is vulnerable to several security threats including potential plots to

place explosives aboard aircraft; illegal shipments of hazardous materials; criminal

activities such as smuggling and theft; and potential hijackings and sabotage by

persons with access to aircraft. Several procedural and technology initiatives to

enhance air cargo security and deter terrorist and criminal threats have been put in

place or are under consideration. Procedural initiatives include industry-wide









consolidation of the "known shipper" program; increased cargo inspections; increased

physical security of air cargo facilities; increased oversight of air cargo operations;

security training for cargo workers; and stricter controls over access to cargo aircraft

and air cargo operations areas. Technology being considered to improve air cargo

security includes tamper-resistant and tamper-evident packaging and containers;

explosive detection systems (EDS) and other cargo screening technologies; blast-

resistant cargo containers and aircraft hardening; and biometric systems for worker

identification and access control (Elias, 2007). While the primary policy focus of

legislation has been on cargo carried aboard passenger aircraft, air cargo security also

presents a challenge for all-cargo operators (FAA, 2006).

History shows that very few accidents were caused by hazardous cargo content.

According to the National Transportation Safety Board's aviation accident database

(NTSB, 2010), in the last 15 years, less than 20 accidents occurred from that cause,

which corresponds to less than 0.1% of all accidents and incidents within that period of

time. Moreover, only one major (fatal) accident was caused by fire in a cargo hold.

Fortunately, the low numbers of accidents do not slow down the aviation authorities in

encouraging the design of safer aircraft.

Fire detection

For some aircraft compartments a fire / smoke detection system is required by the

regulations JAR (Joint Aviation Requirements) and/or FAA. In addition, aircraft

manufacturers install supplementary fire / smoke detection systems to increase the level

of safety. These systems must comply e.g. with the regulations (Schmoetzer, 2001).

The urgency of the corrective action subsequent to a fire / smoke warning depends

directly on the risk and is reflected in the procedures to be applied by cockpit or cabin









crew. For example, a cargo compartment smoke warning is indicated to the flight deck

crew as a red warning, this means the crew has to perform the action immediately. As

long as the crew is unable to differentiate between a true and a false warning, it has to

follow the certified procedure. The impact of a false fire / smoke warning in non

accessible compartments is extensive and might include: flight diversion, declaration of

emergency situation, eventually passenger evacuation, compartment inspection, fire

extinguisher replacement, passenger disappointment, loss of confidence in the warning

system, etc (Schmoetzer, 2001).

Technologies

Because the capability of available technology is seen as a significant constraining

factor on the ability to screen, inspect, and track cargo, initiatives to improve cargo

screening technology have been a focus of recent legislation to enhance air cargo

security. Various technologies are under consideration for enhancing the security of air

cargo operations, such as:

* Tamper-evident and tamper resistant packaging and container seals
* Cargo screening technology using x-rays, chemical trace detection systems, or
possibly neutron beams
* Canine teams
* Hardened cargo container technology
* Biometric technologies

In addition, technologies to better track cargo shipments are being considered to

maintain better control and tracking of cargo shipments along the supply chain. Both

global positioning system (GPS) and radio-frequency identification (RFID) technologies

are seen as emerging technologies for improving the tracking of air cargo in the supply

chain (Elias, 2007). And within that supply chain, there is a definite weak point during air

transit. Therefore, there is a growing interest to know which cargo is on board, where is









it in the cargo hold and what is its temperature, humidity, acceleration, etc (Schmoetzer,

2005).

RFID in Aviation

Cerino and Walsh (2000) stated that RFID technologies and systems with potential

application to the worldwide aviation industry will most likely operate in the international

Industrial, Scientific and Medical (ISM) frequency spectrum. Operation in an ISM band

has the distinct advantage of not requiring the user to obtain specific licenses on a site-

by-site basis. In fact, only an "honor system" compliance with the defined ISM

regulations on a band-by-band and geographical region-by geographical region is

required to allow completely unrestricted site operation anywhere within that ISM region.

In addition, for the most part, the application of RFID for aviation has focused on

the "electronic baggage tag" as a replacement for a barcode baggage tag. However, the

FAA and others have also addressed issues such as RFID for use with baggage

containers, passenger and cargo tracking and as such have considered other additional

frequencies; one such is the 915MHz ISM band (Cerino and Walsh, 2000). The

International Air Transport Association (IATA) member airlines unanimously approved

the IATA Recommended Practice (RP) 1740C document, which endorses the use of

ultra-high frequency tags and readers compliant with the EPCglobal Gen2 protocol as a

global air interface standard for RFID baggage tags (O'Connor, 2005).

ISM Frequency and Aviation RFID Considerations

The US Federal Aviation Administration (FAA) to a large extent, and other aviation

industry vendors and / or air carriers to a lesser extent, has completed testing of various

technologies to ascertain their potential performance for aviation operational utility.

Each system (being comprised of reader and tag) was not tested side-by-side, nor









under identical simulated or operational conditions, yet a rather extensive matrix of

frequency (ISM frequencies of 125kHz / 132kHz, 13.56MHz and 2.45GHz) versus

operational performance was obtained as result of the totality of FAA trials. In total,

seven test phases geared at addressing the full range of aviation functional

requirements along with other additional site-specific operational RFID usage

evaluations contributed to the following results.

Cerino and Walsh (2000) analyzed the results of three years of testing and

research to develop an effective RFID system for the airline/airport environment. It was

found that the 2.45GHz system has better performance, is more flexible in design, can

be assembled off-the-shelf, and the system and tags (about 1/3 the cost of 13.56MHz

disposable tags) are least expensive. The 13.56MHz system requires a mostly

customized design, is less mature than the 2.45GHz technology, and interference

concerns at 13.56MHz add significant complexity and cost to the system. The system at

125kHz / 132kHz presents significant tag cost disadvantages, which are not likely to be

overcome for aviation use.

It is important to recognize that for each ISM band there exists specific transmitter

power, signal modulation, duty cycle and other technical parameters that effectively

comprise the band's operating regulation. These regulations are not consistent from

band-to-band and (coupled with the natural differences in propagation characteristics for

each band) result in significant differences as they relate to efficient aviation industry

use.

2.45GHz: At this frequency the communication is entirely propagation coupling.

Propagation in this band is via directional antennas, and hence reader energy can be









directed to the area of greatest tag likelihood. The 2.45GHz band essentially evolved

with the understanding that ISM interference exists at every band, and hence offers

several advanced communications protocols which counter the potential interference

effects.

Aviation Applications

Although the FAA's interest in RFID stemmed from mandates associated with the

Vice President's Commission on Anti-Terrorism, in particular its application to RF

baggage tags and positive passenger bag matching, the use of RFID for commercial

aviation extend beyond one security application. The fact is, there are many business

and security reasons for applying RFID to the airport environment. Communications and

RFID technology programs represent the key to integrating all the components of airport

security, including perimeter intrusion detection, personnel screening, checkpoint

screening, vehicle and cargo screening, digital video surveillance and recording, and

RFID baggage and vehicle tracking (Hallowell and Jankowski, 2005).

Passenger baggage sortation

In aviation industry, major airports have been looking for opportunities to use RFID

technologies in baggage handling areas since 1999. Many pilot tests have been done at

numerous airports including Gimpo (Korea), Las Vegas, Jacksonville, Seattle, Los

Angeles, San Francisco, Boston, New York, Heathrow and Rome (Chang et al, 2006).

Ouyang et al. (2008) presented an intelligent RFID reader and its application in the

airport baggage handling system. Jacinto et al. (2009) presented an RFID equipment

tuning and configuration methodology developed in a project to support baggage

tracking and feed dashboards with real time status of Service Level Agreements

between the airport, the airliner and the ground operators.









A survey of the aviation industry would quickly identify that most air carriers utilize

optical barcode "baggage tags" to identify the travel itinerary for each individual

baggage item. The aviation industry is working towards standardization of an RF

baggage tag, in hopes of eliminating current barcode read-rate limitations. These

limitations become particularly evident with transfer baggage where the use of

automated sortation systems is essential to keep connection times to a minimum. The

FAA has, with the cooperation of many vendors, airlines and airports, clearly

demonstrated the ability to utilize RF baggage tags to enhance baggage sortation

(Cerino and Walsh, 2000; Chang et al., 2006).

Passenger/ baggage matching. This process necessitates insuring that only

boarded passenger's baggage is loaded onto the aircraft. This process today can be

accomplished either totally manually or semi-automatic (using handheld barcode).

Either way, a significant amount of baggage handling is required and as such there

exists a dependence on human accuracy and barcode quality. The later becomes

particularly suspect with transfer baggage where the tags are no longer in the "pristine"

condition as when originally issued. In addition, certain low-cost, high-performance

RFID tags afford the opportunity to "add to the tag" more than just the IATA license

plate data. Information such as: the results of security screening; passenger biometrics;

baggage images; and, flow timing through the baggage handling process all provide for

significant core business and security benefits (Cerino and Walsh, 2000).

Verification / authentication

From a security standpoint, few processes are considered complete without

featuring a verification / authentication element. For example, if a certain passenger-

checked bag is screened and considered to be cleared, this "cleared status" may be









used as basis for loading on to an aircraft. Consequently, it is imperative that the overall

process be able to accurately verify that the bag in question definitely is the bag that

has been deemed to be cleared for loading versus any other bag in the system. A

robust RFID system implementation could easily track that bag based on its physical

characteristics (with the support of other sensors); unique data (securely) added to the

tag and/or overall RFID carrier IT system; and, know path / timing for the bag through

the baggage handling process (Cerino and Walsh, 2000). McCoy et al. (2005)

investigated an automatic tacking system to improve airport security and efficiency by

means of a cellular network of passive RFID receivers, combined with far-field active

RFID tags which may be issued within boarding cards or as security badges.

Tracking and locating

The tracking and rapid locating of baggage, cargo and containers (and the

associated integrity assurance of those items) is essential to the overall security of a

commercial flight. This tracking / locating of "transport items" is accomplished today, for

the most part, only by a very labor-intensive manual process. The application of RFID

not only would be more time and labor efficient, but also more accurate. Sensors and

electronic systems do not fall asleep, forget their assignments, become distracted, or

otherwise perform in an unpredictable manner. The introduction of RFID to provide this

asset tracking and locating offers the opportunity for: centralized monitoring; continuous

surveying; automatic event logging; and, of course, more rapid finding of items when

retrieval is mandatory (Cerino and Walsh, 2000).

Boeing and Airbus are also promoting the adoption of industry solutions for RFID

on commercial aircraft parts. They believe that RFID could provide major benefits for

the entire industry. They will get more accurate information about their demand for parts









and will be able to reduce their parts inventory and cut the time it takes to repair planes.

Part suppliers will also be able to reduce inventory, improve the efficiency of their

manufacturing operations and reducing the amount of unapproved parts that enter the

supply chain (Chang et al., 2006).

Cargo

A natural evolution, beyond the sortation of passenger baggage, is the use of RF

tags to handle sortation of cargo. For this application, the process is quite similar to

passenger baggage sortation, with exception that cargo parcels often can have a rather

large form factor. As such, only an RFID system with enough flexibility and performance

features to allow achievement of cargo and passenger baggage sortation requirements

would be a realistic option for most of the aviation community. Naturally, for carriers

such as FedEx and UPS this restriction do not directly apply, however, even in those

examples, there is benefit in commonality with passenger air carrier systems. With

cargo, however, the concept of "tagging" entire palletized shipments becomes a

consideration. This could demand fixed station RFID readers requiring an even greater

read volume or, more practically, implementation of an area read capability (such as the

entire cargo hanger or even the entire tarmac area) to cover widely dispersed assets.

This concept of broad coverage area for very large parcels also relates to passenger

baggage, as in containerized luggage for stowage on wide-body aircraft. And, as it

relates to efforts accomplished and planned by the FAA, the application of RFID to

ensure the integrity (from a security standpoint) of ULDs from the time they are filled

until the time they are loaded onto the aircraft (Cerino and Walsh, 2000).

Containers. According to Cerino and Walsh (2000), passenger baggage

containers (ULDs), are an important aspect of commercial aviation. For wide-body









aircraft, they are used to store up to 70 individual pieces of luggage (depending, of

course, on the size of the individual pieces) and are loaded as a single unit. Whenever

possible, baggage handlers segment passenger baggage by destination, in an effort to

reduce the amount of handling time for luggage. This process of sorting luggage by

destination and container positioning, besides supporting minimized ground turn-around

time, is important to support the security process of passenger bag matching.

Accordingly, the use of RFID to: manifest exactly which baggage is in which container;

support the matching of the baggage to passenger "aboard aircraft" status; and exactly

locate the container which holds the passenger baggage, in valuable from both security

and operational efficiency standpoints.

Monitoring. With air freight increasing rapidly; just in time delivery being a real

challenge; ULDs being packed by third parties; inspection time being very limited and

transportation security being linked to volume, monitoring of goods within containers

may contribute to security. Moreover, customers and insurance providers want to know

what happens to their shipment (liability issue) and forwarders need to increase the

monitoring of goods. Improvements may be feasible for temperature sensitive goods,

hazardous materials and high value cargo, and such monitoring require communication

between the container and the aircraft (Schmoetzer, 2005). If readers are installed in

the aircraft, they can be used to interrogate the cargo, acquire data from dedicated

goods, enhance the monitoring, set off warnings, etc (Figure 2-7). Therefore, automated

onboard identification of cargo can contribute to enhance of security (Schmoetzer,

2005). Benefits of RFID assisted air freight handling include:

* Automated tracing of goods
* Automated verification of aircraft load instruction









* Reduction of false loading
* Reduction of ground time
* Paperless data transfer
* Electronic Bill of Loading
* Wireless interface can be used to enable more services than simply RF-
Identification e.g.:
o Change / update information on relevant item
o Self control / monitoring means
o Memorize what is of interest
o Data exchange (e.g. actual temperature, history, etc.)

Cold chain

Products, such as food, pharmaceuticals and flowers, are at high risk of perishing

from various adversities along the cold chain. The parties involved should control when

possible, and at the very least monitor the conditions of the goods in order to ensure

their quality and to comply with all legal requirements. Among environmental

parameters during transport, temperature is the most important in maintaining the shelf

life of the products (Nunes et al., 2006; Zweig, 2006; Jedermann et al., 2009).

With today's globalization, there is a growing need for fresh products to be

delivered year round all over the world, thus, temperature sensitive items are likely to be

shipped by air because of their relatively short shelf life. Unfortunately, a faster transit

time does not always imply controlled temperature throughout transportation. In

contrast, during airport operations, loading, unloading, air transportation or warehouse

storage, perishable goods often suffer from temperature abuse either due to difficulties

in controlling the temperature, absence of refrigerated facilities, or lack of information

about produce characteristics and needs (Nunes et al., 2003). Of approximately 2.6

million tons of perishables air freighted in 2008, nearly 30% is estimated to be lost due

to handling and temperature abuse (Catto-Smith, 2006). In a previous study, Emond et

al. (1999) showed that the environmental conditions during airport operations could, in









fact, be very far from the optimum for fruits and vegetables. Moreover, in a strawberry

quality study, Nunes et al. (2003) showed that greater losses in quality occurred during

simulation of the airport handling operations, in-flight, and retail display than during

warehouse storage at the grower, truck transportation to or from the airport, or during

backroom storage at the supermarket. The relative success of growing exotic

perishables in countries such as Colombia, Ecuador, and Peru in South America, and

its successful distribution and commercialization in distant markets such as the US and

the EU, have been made possible due to advances in transportation and refrigeration

technologies. Yet the transportation systems of exotic perishables are far from perfect.

Transportation and logistics costs can be high both monetarily and in terms of loss of

quality during handling (Vega, 2008).

While passenger business is generally bidirectional, cargo is not. Consequently,

freighter routes are often imbalanced. This implies that when transporting goods from

point A to point B, the freight rate charged must also cover the return trip from B to A

(Vega, 2008).

Temperature monitoring. Currently, most digital temperature loggers have to be

connected to a host device to download data, and as a result, have limited real-time

data interactivity, which results in after-the-fact analysis for claims, loss in quality and

related issues. Radio frequency identification (RFID) temperature loggers function

wirelessly which allows for real-time information transfer. Active or semi-passive RFID

tags can support one or many sensors as well as the unique ID that RFID technology

provides by design. The RFID tag, with associated hardware and software adds the

benefit of having the item scanned on receipt, so that if an alert (alert triggers are









programmable prior to shipping) is active, the receiver knows immediately (not after-the-

fact) that there is a potential problem with the shipment and can spend the time required

on specific shipments rather than going through random inspections (Jedermann et al.,

2007). Many studies have already shown the effectiveness of RFID in monitoring

product temperature during transit (Emond, 2007; Jedermann and Lang, 2007;

Jedermann et al., 2007; Ketzenberg and Bloemhof-Ruwaard, 2009).

Wireless Interference

Portable electronic devices' (PEDs) interference risk to aircraft radio systems is a

concern during flights. For various reasons, many devices such as laptop computers are

allowed during flights, while intentional transmitters such as wireless devices and

phones are prohibited. There have been many past studies addressing the three

elements. Examples include emissions measurements from wireless devices in aircraft

radio bands (Nguyen et al., 2004, 2005). Aircraft radio receiver interference thresholds

data may be found in "Environmental Conditions and Test Procedures for Airborne

Equipment" (DO-160F) published by the Radio Technical Commission for Aeronautics

(RTCA).

Devereux et al. (1997b) reported research and experimental results of commercial

aircraft avionics and control systems exposed to conducted and radiated

electromagnetic interference. This experiment, conducted inside a CV-580 aircraft,

determined how susceptible installed avionics are to different low power RF sources

located in passenger cabin and baggage compartments, and avionics and cargo bay

areas. This study showed that avionics certified to the special 100V/m high intensity

radiated field requirements are still highly vulnerable to low to moderate levels of

onboard RF energy when installed in the aircraft system. Various sources of









electromagnetic interference can be easily transmitted at these frequencies. Also, many

of the avionics bays and cable routes onboard today's modern aircraft are unshielded

and easily accessible. Many of the aircraft cargo bays are located adjacent to the

avionics bays with only fiberglass walls and access doors separating the two (Cerino et

al., 1997). A cargo hold could easily contain a RF transmitter system which emits

enough RF energy to seriously interfere with and upset today's avionics systems.

Further analysis of the CV-580 aircraft was made by Devereux et al. (1997a) which

emphasizes the aircraft receiver in-band frequency susceptibility, primarily the

microwave frequency bands used for aircraft Global Positioning Satellites (GPS) and

Satellite Communication (SATCOM) navigation. This report also includes VOR, VHF,

UHF and DME navigation and communication frequency bands (refer to Table 2-3). The

evaluation is focused on coupling behavior from inside the aircraft cavity, coupling

through the windows, to the aircraft receive antennas. It was found that path loss

coming from the cargo bay is likely to travel up through the mostly non metallic ceiling of

the cargo bay into the passenger cabin and out the passenger windows or through the

non metallic avionics bay doors located in the cargo bay through to the cockpit window

and to the receiver antennas. Moreover, the path losses from a transmitting antenna at

a central position inside the fuselage were greater than that from any window and about

10 to 20 dB greater than from the optimum window.

Furthermore, in a study by Nguyen et al. (2007) interference coupling factor (or

interference path loss) data were measured for multiple radio systems on ten small

aircraft. The data show significant data variations between different aircraft models. The









data also show stronger interference coupling than for previously measured larger

aircraft, and potentially result in higher interference risks.

Electronic devices

Unlike aircraft installed equipment, passenger carry-on devices such as wireless

phones are not required to pass the rigorous aircraft radiated field emission limits.

Previous studies were made to measure the emissions from wireless phones in aviation

bands and to assess interference risks to aircraft systems (Ely et al., 2003; Nguyen et

al., 2004). Results from a recent study showed that the spurious emissions from 33

phones tested were below the aircraft installed equipment limits (RTCA/DO-160 cat. M),

even with the consideration of the 5-8dB uncertainty associated with the phones

expected directivity (Nguyen et al., 2005).

RFID interference

Passive tags are considered less of an interference concern for aircraft since they

do not transmit without an interrogator, whose electromagnetic fields power the tags.

Active tags, on the other hand, are powered with internal batteries. Active tags can be of

higher interference risks since many can transmit on their own without an interrogator.

The actual interference risks depend on several factors, including the tags' intentional

and unintentional emission levels, the propagation path loss factor, and the victim

system's susceptibility threshold to the emissions type. Nguyen et al. (2008) studied the

emission measurements of active tags and their interference potential on aircraft

sensitive radio receivers. Specifically, this study measured the unintentional emissions

from several popular RFID tags used for cargo tracking. Results showed that many

tags' peak total radiated power exceeded RTCA/DO-160E category L and M EIRP

emission limits, one of which surpassed the limit by as much as 35 dB in the GS band









(328.60 335.40MHz). Another study made with two active tags at 433MHz showed

that emissions from both tags were higher than the limit of DO-160E at the operation

frequency and the harmonics. It was also found that the emission level depends greatly

on the location and material on which a tag is placed (Yonemoto et al., 2007).

For a complete interference assessment, other factors such interference path loss

and receiver interference thresholds should also be considered. These factors were

addressed previously, for example, Nguyen et al. (2006) documented the measurement

of interference path loss for cargo bays on a Boeing 747 and an Airbus A320 aircraft.

Nguyen et al. (2004) provided a summary of passenger cabin path loss data for many

commercial transport aircraft. Nguyen et al. (2007) reported the path loss

measurements for general aviation aircraft. These path loss data represent the

propagation loss between the tag locations and the victim receiver's antenna port.

Aircraft radio receiver interference thresholds for continuous interference signal

transmission were addressed in RTCA/DO-294B (Guidance on allowing transmitting

portable electronic devices (T-PEDs) on aircraft). LaBerge (2007) analytically

determined thresholds for intermittent interference signals similar to RFID emissions.

Moreover, the work in Nguyen and Mielnik (2008) reports the laboratory effort to

determine the GS system interference threshold to an RFID interference signal.

RFID airworthiness policy

The current Federal Aviation Administration (FAA) RFID policy document is AC20-

162 "Airworthiness approval and operational allowance of RFID systems" (FAA, 2008).

It incorporates and supersedes jointly-issued AIR-100, AFS-200 and AFS-300 RFID

policy. This advisory circular offers guidance on installing and using RFID systems on

aviation products and equipment. Specifically, it provides an acceptable way to use









RFID readers or interrogators installed on aircraft, and advice on allowing use of RFID

devices on baggage, mail containers, cargo devices and galley / service carts. It also

covers using portable RFID readers or interrogators carried onboard aircraft. This

advisory circular is not mandatory and does not constitute a regulation. It describes an

acceptable means, though it is not the only means, to show compliance with applicable

installation and operational requirements. Through this AC, the FAA does not prohibit

any use of RFID devices, it provides specific requirements that the equipment must

meet to be safe and airworthy.

Concluding Remarks

This review showed that RFID has long been thought to have the potential to help

the air cargo industry by increasing its safety, operation efficiency, monitoring capability,

customer satisfaction, etc. Before implementation can take place, some questions have

to be answered and more research has to be done. This is why this study will cover the

following subjects: interference and frequency assessment in air cargo warehouses; RF

propagation study inside the cargo hold; RF propagation through ULD materials; and

temperature distribution in the cargo hold during transit.









Table 2-1. EPCglobal tag class structure
EPC Class Functionality Type
Class 0 Gen 1 Read only Passive
Class 1 Gen 1 Write once, read many Passive
Class 1 Gen 2 Write many, read many Passive
Class 2 Write once, read many Passive
Class 3 Read and write Semi-passive
Class 4 Read and write Active
Class 5 Reader tags Active

Table 2-2. Permissible field strengths for RFID systems in accordance with FCC Part
15 (FCC, 2008).
Frequency range Max. E field Measuring distance
(MHz) (mV/m) (m)
13.553-13.567 10 30
433.5-434.5 11 3
902.0-928.0 50 3
2435-2465 50 3
5785-5815 50 3


Table 2-3. Aircraft
Aircraft Band
LOC
VOR

VHF-Com

GS
DME
ATC
TCAS/TCAD

GPS (L5)
GPS (L2)
GPS (L1)
MLS


radio systems
Description
Localizer
Very high frequency omnidirectional
range
Very high frequency voice
communication
Glideslope
Distance measuring equipment
Air Traffic control radar beacon system
Traffic collision avoidance system /
Traffic collision alert device
Global positioning system
Global positioning system
Global positioning system
Microwave landing system


Receive Spectrum (MHz)
108.10-111.95
108.00 117.95

118.00 138.00

328.60 335.40
962.00 1213.00
1030.00
1090.00

1176.45
1227.60
1575.00 2
5031.00 5090.70











c 0 0C 0

N N N N Electromagnetic

T a Spectrum



2, / O O O O O Frequency(Hz)



< g Ca
3) 3 5- 3 0
1 1 ) II





3 7- 3 3

I I

S 8 o 1
3 3






S3 Waveln!h Wavelength
3 3 3 3

Figure 2-1. Electromagnetic spectrum










Am plilCe Amrplitude
(Positve I A a Certain Point
Am pliide)




Anpllrde
(Negarve
Amplilude ,





Waveiength




64








Figure 2-2. Different parts of a wave (Lahiri, 2006)




U. __.))

a RFID Reader







Figure 2-3. Example of an RFID system on conveyor belt.


Linear Polarized


x
7 goal96
ei


Figure 2-4. Wave propagation for linear and circular polarization


Antenna


Circular Polarized


7


Y









Airside (import)
m


I..


Airside (export)
en


Breezeway


Break-down


FS Storage


Import Export


SLandside(import)


Storage




andside (export) : .


Figure 2-5. Cargo warehouse floor plan and activity areas.


AS


Ui


Build-up


--1ri -










iipmfl*M i: -4


START


U-


STCPOVER
. ---- PINIUN -----




r


*
ac f


SM za10 I D In 14 1)On
I igh-(>29AMC) or ow<-(<72*C) tempera.


Figure 2-6. International flight temperature profile with both high-temperature
excursions (during stopovers) and low-temperature excursions (in flight)
(Syversen et al., 2008).


COniUMwID AFA23-U- ------ --
Caniner Type AFA
Corner SMait Loaded
PAcess Saus CbHed


Figure 2-7. Example of a GSM/GPS capable RFID system for real time data acquisition
(Schmoetzer, 2005).









CHAPTER 3
AIR CARGO WAREHOUSE ENVIRONMENT AND RF INTERFERENCE

Introduction

The air cargo system is a complex network that handles a vast amount of freight,

packages, and mail carried aboard passenger and all-cargo aircraft. With air freight

increasing rapidly; just in time delivery being a real challenge; unit load devices (ULDs)

being packed by third parties; inspection time being very limited and transportation

security being linked to volume, monitoring of goods within containers may contribute to

enhanced security (Schmoetzer, 2005). Moreover, monitoring the movement of goods

within the air cargo warehouse can enhance operation efficiency and provide valuable

real-time information.

Cargo warehouse. A typical cargo terminal is divided into an import area and an

export area. The import area is dedicated to receiving, processing and releasing

inbound freights. The export area is dedicated to receiving, processing and preparing

outbound freights. The flow of goods through the terminal is either from the airside to

the landside (terminating freights or connecting freights requiring the road feed service),

from the landside to the airside (originating freights or connecting freights arriving from a

road feeder service), or from the airside to the airside via the terminal (connecting

freights). These movements of goods in many directions can easily lead to item

misplacement, and therefore it can be time consuming for the air cargo agents to locate

the pieces when they are needed for build-up. Item level radio frequency identification

(RFID) tagging, as well as ULD tagging, can address this issue by allowing real-time

access to location of all shipments present in the warehouse. Moreover, RFID could

permit automated and efficient ULD / item association during container build-up. This









technology could also give time stamp data on the movement of ULDs in and out of the

building. There are many applications for which RFID could be helpful in an air cargo

warehouse, but for a reliable system to be working properly, interference assessment

needs to be done.

RF interference. Interference is typically interpreted as other sources of radio

frequency (RF) energy that compete with an existing implementation. Any devices

operating at the same frequency have the potential to interfere with each other. There

exist many short range consumer devices operating at UHF frequencies. For example,

devices like automotive remotes, alarm systems, home automation and wireless

temperature sensors, can operate at 433MHz. At 915MHz, GSM 900 is a potential

threat, as well as cordless phones, stereo, older wireless local area network (WLAN),

amateur radio, etc. At 2.45GHz, interference is possible from devices such as newer

WLANs, cordless phones, microwave ovens, fluorescent lighting and Bluetooth

technology to name a few. There are ways to work around this issue by setting up these

systems so that they do not use the same RF channels at the same time, but their

consideration as potential interference threat to any UHF RFID system is of prime

importance.

Frequency allocation. RFID is a radio technology and as such requires the use of

radio spectrum to operate. Generally, when discussing spectrum issues, focus tends to

be on tags that make use of the ultra high frequency (UHF) range. UHF (together with

VHF) is the most common frequency band for television. In addition, it is used for mobile

telephony, two-way radio communication and increasingly for digital services. Since

RFID shares the UHF range with other applications, only a limited bandwidth within









UHF is available for short range devices and RFID, which is 902-928 MHz in North

America. Although there are RFID tags and applications that make use of other

frequency bands (for example 433MHz and 2.45GHz), most spectrum issues seem to

concentrate around the UHF range and many sectors and applications (e.g. in retail and

supply chains) use UHF tags (van de Voort and Ligtvoet, 2006).

ISM band. The Industrial, Scientific and Medical (ISM) bands are defined by the

International Telecommunication Union Recommendation (ITU-R) in 5.138, 5.150, and

5.280 of the Radio Regulations. Individual countries' use of the bands designated in

these sections may differ due to variations in national radio regulations. Because

communication devices using the ISM bands must tolerate any interference from ISM

equipment, these bands are typically given over to uses intended for unlicensed

operation, since unlicensed operation typically needs to be tolerant of interference from

other devices anyway. In the US, the use of ISM bands is governed by Part 18 of the

FCC rules, while Part 15 Subpart B contains the rules for unlicensed communication

devices, even those that use the ISM frequencies. All frequencies considered in this

study are part of the following ISM bands: 433.05-434.79MHz, 902-928MHz (North

America only) and 2400-2500MHz.

Spectrum analyzer. The sensitivity (or threshold) of a spectrum analyzer is

defined as its ability to detect signals of low amplitude. The maximum sensitivity of the

spectrum analyzer is limited by the noise generated internally (Anritsu, 2008). Noise

level is directly proportional to the resolution bandwidth (RBW) of the system. Therefore,

by decreasing the bandwidth by an order of 10dBm on the logarithmic scale (from

100KHz to 10KHz, for instance), the system noise floor is also decreased by 10dBm. As









an additional example, when the RBW is increased from 100Hz to 10kHz, the noise

floor moves up 20dBm (Figure 3-3). When comparing spectrum analyzer specifications

it is important that sensitivity is compared for equal bandwidths since noise varies with

bandwidth. For this test, all data were taken at 10kHz RBW, which was the maximum

sensitivity achievable with this specific spectrum analyzer. In other words, the real noise

floor could have been lower than what was recorded.

Reader sensitivity. RFID communication goes two ways, first the reader-tag link,

then the tag-reader link. The reader sends an RF signal with initial output power, which

attenuates while traveling in the medium between the reader and the tag. If the tag

received enough signal to respond, it sends its info back to the reader, which is also

attenuated on the way back. When the reader receives this much attenuated signal, if it

is above reader sensitivity levels, the communication is complete. The reader sensitivity

defines the minimum signal level needed to be able to communicate with the

transmitter. For passive RFID, it is usually the reader-tag link that is the limiting one,

whereas for a semi-passive tag, the reader-tag link requirement is much more lenient

since the received power must only be decoded not exploited (Dobkin, 2008). Reader

sensitivity is dependent on several design choices, particularly, antenna configuration,

and will become more important as tag IC power is scaled to lower values. Moreover,

when interference is present, it is the tag-reader link or reader sensitivity that limits the

system the most. The sensitivity of the receiver is limited by the noise that enters it and

the largest source of noise for an RFID receiver is usually the leakage from its own

transmitted signal (Dobkin, 2008). Nevertheless, any other sources of interference,

within the specific UHF band used, can affect the performance of an RFID system when









its noise level is above the reader sensitivity limit. For example, a few years ago the

typical passive RFID reader had a sensitivity of -65dBm, whereas in 2007 the most

sensitive reader had a sensitivity of -77dBm (Impinj, 2007), but today Intelleflex's new

XC3 reader has a receive sensitivity of -120dBm (Intelleflex, 2010). This does not mean

that any noise above -120dBm would completely disable the communication link

between a tag and reader. Interference can only be expected in the case where the

noise level would be well above the tag-reader link; and that is not taking into account

the fact that most RFID readers have their own interference filtering system. In other

words, each RFID reader will tolerate up to a certain level of signal-to-noise-ratio (SNR

in dB) which indicates how much higher the signal level is compared to the noise level.

For instance, if a reader with sensitivity Y dB requires an SNR of X dB for a successful

read, this means that in order to use the reader at its maximum capacity, the noise level

present in the environment has to be less than (Y X) dB. If this is not the case, than

the effective distance between the tags and the readers need to be adjusted accordingly

to compensate for the environmental noise.

For RFID technology to be implemented into an air cargo warehouse, many RF

assessments need to be done. In particular, interference levels might dictate which

frequency is most suitable for a specific RFID application. Therefore, the objective of

this study was to identify the multiple RF interference encountered inside air cargo

warehouses.

Materials and Methods

RF interference readings were taken at different locations throughout two air cargo

warehouses. One warehouse was located near the Pierre-Elliott Trudeau airport (YUL)

in Montreal, Canada, whereas the other one was near Pearson airport (YYZ) in Toronto,









Canada. All interference readings were taken with a handheld spectrum analyzer (HSA

9101, Willtek, Parsippany, NJ) at three different UHF frequencies: 433MHz, 915MHz

and 2.45GHz. Each frequency was measured using a specific receiving antenna, as

detailed in Table 3-1. Signal data were adjusted to account for receiving antenna gain.

In other words, readings at 915MHz were lowered by 2.5dBm and readings at 2.45GHz

were lowered by 8dBm.

All interference data are in dBm and the range considered for each frequency is in

accordance with the Federal Communication Commission part 15 regulations (FCC,

2008). For 433MHz, the range of device operation is between 433.5 and 434.5MHz; for

915MHz, the range is between 902 and 928MHz; and for 2.45GHz, the range is from

2.435 to 2.465GHz.

Export: All freight arrives at the cargo facility either as bulk or as shipper loaded

unit devices (SLUD), which can be air containers or pallets, pre-loaded by the customer.

The freight gets weighed and dimensioned by the acceptance agent and stored at the

appropriate location depending on its flying time and destination. If items are bulk, they

ultimately go to the build-up area to be put in a ULD (pallet or container) or are

transported in a tub cart directly to the airplane if this airplane is bulk loaded. ULDs are

transported onto roller systems or forklifts through the cargo facility, after which they are

carried out to the ramp onto trailers.

Import: ULDs are transported between the aircraft and cargo facility on special

trailers pulled in a train by small tractor vehicles. Bulk items are unloaded from the

aircraft directly into tub carts. Everything is brought back to the cargo warehouse and is

broken-down when needed and stored until customer pick-up.









Montreal Warehouse

This location is in a very new building (constructed in 2008) and covers an area of

16,300m2. Figure 3-1 presents the layout of the warehouse, including most activity

areas (export, import, storage, build-up, break down, airside and landside). Interference

reading points are circled on the layout and correspond to the following areas:

* Office area
* Export landside, near unloading dock doors
* Breezeway, export airside, near doors
* Breezeway, import airside, near doors
* Center of warehouse, between import and export; pallet break-down and build-up
area; and storage
* Import landside, near loading dock doors

Toronto Warehouse

This location is slightly older building, built in 2002, and covers an area of

26,700m2, which is more than 60% bigger than the Montreal warehouse. Figure 3-2

presents the layout of the warehouse, including most activity areas (export, import,

storage, build-up, break down, airside and landside). Interference reading points are

circled on the layout and correspond to the following areas:

* Import landside, near loading dock doors
* Office area
* Export landside, near unloading dock doors
* Build-up area
* Center of warehouse, between import and export
* Break-down of imported ULDs
* Breezeway, import airside
* Middle of breezeway
* Breezeway, export airside

Results and Discussion

As stated earlier, frequency ranges are as follows: 433.5 to 434.5MHz; 902 to

928MHz; and 2.435 to 2.465GHz. Table 3-2 and 3-3 show minimum and maximum









electromagnetic signal levels recorded inside these ranges for Montreal and Toronto

respectively. The "maximum" data point is the worst interference level recorded in the

area (Table 3-2 and 3-3). The "minimum" data is only present to show a comparison

base for peak signal identification. It represents the sensitivity of the spectrum analyzer.

When minimum and maximum data points are similar for a single location, it can be

concluded that no "peak" is present and therefore no specific interference was observed

above noise floor.

433MHz

Around 433MHz, the study did not show any important signal peaks that could

interfere with an RFID system operating in this range. The average minimum

background noise is -107.3dBm and the maximum signal level recorded is -103.7dBm in

Montreal for position 1 and 3 (Table 3-2); and -89.8dBm in Toronto, position 4 (Table 3-

3). Figure 3-4 shows the spectrum analyzer graphs for 433MHz in Toronto, position 4.

Only one graph per frequency has been chosen as illustration, which represents the

worst case scenario recorded within all locations and positions at that frequency.

According to the National Telecommunications & Information Administration (NTIA,

2003, 2010) the small signal peaks visible around 433MHz (Figure 3-4) corresponds to

"RADIOLOCATION and Amateur" (Capital letters are primary activity and lower cases

are secondary) in the range of 420-450MHz.

The ISM range 433.05-434.79MHz can be heavily occupied by a wide range of

ISM applications. In addition to backscatter (RFID) systems, baby intercoms, telemetry

transmitters (including those for domestic applications, e.g. wireless external

thermometers), cordless headphones, unregistered LPD walkie-talkies for short range

radio, keyless entry systems (handheld transmitters for vehicle central locking) and









many other applications are crammed into this frequency range. Unfortunately, mutual

interference between the wide range of ISM applications is not uncommon at this

frequency (Finkenzeller, 2003). However, as far as the results found in this experiment

are concerned, the very minimal peak recorded as the worst case scenario does not

cover the entire band and is not present at a high power level. It is in fact lower than

reader sensitivity of RFID readers designed three years ago. Therefore, it would be safe

to say that the level of interference recorded would not significantly affect the

performance of an RFID system.

915MHz

As shown in Figure 3-5, there is very low activity within the 902-928MHz range in

the Toronto warehouse, which represents again the worst case scenario recorded. The

maximum signal level recorded was -99.8dBm for Montreal and -92.2dBm for Toronto,

whereas the average minimum is -109.5dBm (Table 3-2 and 3-3). As explained earlier,

this level of noise could possibly only interfere with state of the art readers that have

sensitivities below -90dBm and are being used at their maximum capacity (maximum

range which allows for successful communication). In the event that this level of noise

would not be blocked or filtered by the reader, it would still only interfere with tags that

are further away from the reader, and require this much sensitivity to communicate.

Typically, when the interference level is close to the reader sensitivity, it can only affect

the tags that have communication links "power levels" between the interference noise

and the receiver sensitivity. Moreover, systems working in that frequency range in North

America use frequency hopping techniques; which means that the reader utilizes a slim

portion of the band for a maximum of 0.4 seconds at a time. In other words, if the









interfering signal is only present on a small portion of the band, the likelihood of

interference becomes lesser.

As seen in Figure 3-6, activity surrounding this band corresponds to (NTIA, 2003,

2010), where capital letters are primary activity:

* 851-894: FIXED + LAND MOBILE (GSM 850)
* 929-932: FIXED + LAND MOBILE
* 932-935: FIXED
* 935-940: FIXED + LAND MOBILE

Wireless applications that operate license-free in the 900 MHz ISM band include

supervisory control and data acquisition, industrial automation, building automation and

control, wireless sensor networks, and consumer devices such as cordless telephones,

wireless speakers and baby monitors. In the U.S., most such systems utilize frequency

hopping spread spectrum over 902-928MHz.

2.45GHz

Around 2.45GHz, the study shows much more activity that could interfere with an

RFID system operating in this range (Figure 3-6). The maximum signal level recorded is

-74.6dBm in Montreal for position 4; and -66.5dBm in Toronto for position 9, whereas

the average minimum is -112.6dBm (Table 3-2 and 3-3). RFID readers working at this

frequency have similar sensitivities as readers for 915MHz and frequency hop within the

specified range. As mentioned earlier, frequency hopping only allows a small portion of

the band to be used at a time; therefore if the interfering signal is only present on a

portion of the band, the likelihood of interference becomes smaller. Signal level of

around -66dBm is well above most reader sensitivity, which could cause unwanted

interference in the event that those readers are not capable of filtering the interfering

signal.









The signal peaks visible around 2.45GHz for Montreal (Figure 3-6) corresponds to

(NTIA, 2003, 2010), where capital letters are primary activity and lower cases are

secondary:

* 2417-2446: Amateur + radiolocation
* 2454-2472: FIXED + MOBILE + radiolocation

In the 2.4GHz ISM band there are several sources of interfering signals, including but

not limited to: microwave ovens, baby monitors, wireless phones, wireless camera,

Bluetooth enabled devices, WLANs, WIFI, and 2 way radios.

Conclusion

This study showed interference levels at three UHF frequencies recorded in two

air cargo warehouses. The interference levels from highest to lowest were at 2.45GHz,

433MHz, and 915MHz respectively. Even in the case where interference is above a

typical reader sensitivity, it is hard to say that these signals will or will not interfere with

and RFID system since every system is designed differently. When the interference is

on the edge of the spectrum, it is possible to reduce the used RF band in order to avoid

such noise. When the interference is more towards the middle or across the entire

band, RFID system design must take this into consideration to filter the noise out. RFID

systems working at 915MHz and 2.45GHz use frequency hopping, therefore, when

noise is only present on a certain part of the band, interference is also only encountered

when the frequency is "hopping" on that part of the band, which should only create

interference for a short period of time (depending on the percentage of the band

covered by the interfering signal). This test was performed inside two warehouses,

which may or may not give a realistic average representation of any air cargo

warehouse in North America. But as far as these locations are concerned, it would be









safe to say that implementation of RFID systems at 915MHz or 433MHz would bring the

best results, interference-wise.











Table 3-1. Receiving antenna specifications
Frequency Polarization Gain Model & Manufacturer
B-368-1, How Tsen Intl. Electronics Metal
433MHz Linear (omni) OdBi Co.,Ltd. Shin Wu Hsiang, Tao Yuan Hsien,
Taiwan
MHz Linear (omni) 2Bi EXR902TN, Laird Technologies,
915MHz Linear (omni) 2.5dBi Schaumburg, IL
Schaumburg, IL
2.45GHz Linear (omni) 8dBi MRN-24008SM3, AntennaWorld, Miami, FL









Table 3-2. Minimum and maximum interference readings (in dBm) for six positions and three frequencies at the Montreal
warehouse.
Frequency Position in the warehouse
(MHz) 1 2 3 4 5 6
min max min max min max min max min max min max
433 -107.3 -103.7 -107.8 -103.8 -106.6 -103.7 -106.7 -104.4 -108 -104.5 -108.1 -105
915 -109.8 -103.8 -109.8 -100.0 -109.0 -101.9 -109.0 -99.8 -108.6 -100.4 -108.6 -105.9
2450 -113.7 -82.9 -113.9 -79.3 -113.1 -76.1 -112.6 -74.6 -112.4 -82.0 -113.7 -88.7

Table 3-3. Minimum and maximum interference readings (in dBm) for nine positions and three frequencies at the Toronto
warehouse.
Frequency Position in the warehouse
(MHz) 1 2 3 4 5
min max min max min max min max min max
433 -106.6 -103.9 -106.4 -104.2 -107.1 -103.8 -107.4 -89.8 -106.7 -104.5
915 -110.5 -99.9 -110.1 -103.7 -108.9 -95.9 -110.1 -98.1 -110.5 -94.4
2450 -112.0 -90.0 -111.8 -91.3 -112.1 -86.5 -112.4 -76.1 -112.0 -80.5


Table 3-3. Continued.
Frequency Position in the warehouse
(MHz) 6 7 8 9
min max min max min max min max
433 -107.8 -104.2 -107.2 -105.3 -107.3 -101.8 -108.2 -103.9
915 -109 -98.4 -110.1 -102.2 -110.1 -92.2 -108.9 -93.4
2450 -112.1 -90.1 -112.4 -69.4 -112.5 -71.6 -112.5 -66.5









Airside (export)
: """"*, *


Breezeway


K$


Break-down
.. ... .. .. .......................... .... ::: ......... ..........


Storage


Import






SLandside (import) ig7,,;


) Build-up




Export
S Storage
t-I. T -


Landside (export)


W4
1. :.k


Figure 3-1. Montreal cargo warehouse facility floor plan and interference reading points
(numbered 1 to 6).


AR
t-


I


Airside (import)
,--- 1


k._.j









Airside (import)


Landside (import) Landside (export)'

Figure 3-2. Toronto cargo warehouse facility floor plan and interference reading points
(numbered 1 to 9).


>AA. ... A. > :...* 4 -^ > 10KHZ RBW
-r ., L.... ... ...- ..- .- -.. -' IKHzRBW 20dBm
10dBm{- o bw
Figure 3-3. Noise floor of spectrum analyzer at three different resolution ba100H ndRBW



Figure 3-3. Noise floor of spectrum analyzer at three different resolution bandwidths.


Airside (export)











Spectum Analyer 9101 1
Raw 10 km


-40



-80



120


433MHz


will'tek


Frequency (MHz)


Figure 3-4. Worst case scenario for signal interference readings around 433MHz
(Toronto, position 4). Span: 10MHz, RBW: 10kHz, attenuation: OdB, gain:
OdBi.


Spectrum analyze 91r oa1*aue l *


E
m
-o

> -80



O) -120


915MHz


will'tek


Frequency (MHz)


Figure 3-5. Worst case scenario for signal interference readings around 915MHz
(Toronto, position 8). Span: 50MHz, RBW: 10kHz, attenuation: OdB, gain:
2.5dBi.











84


433.5-
434.51 Hz


o1o MHIWDIU


San 1000000 MH


5000 MMDU~w











Spectrum Analyer 9101 0


2450MHz


E


> -80



') -120


Frequency (MHz)
Frequency (MHz)


Figure 3-6. Worst case scenario for signal interference readings around 2450MHz
(Toronto, position 9). Span: 50MHz, RBW: 10kHz, attenuation: OdB, gain:
8dBi.


...301 1'5


50CO MHWDiu


San 00000oo MH









CHAPTER 4
RADIO FREQUENCY PROPAGATION INSIDE THE CARGO HOLD OF A DC-10
AIRCRAFT

Introduction

New technologies to better track cargo shipments are responsible for maintaining

improved control and tracking along the supply chain. Both global positioning system

(GPS) and radio frequency identification (RFID) technologies are emerging technologies

for improving the tracking of air cargo in the supply chain (Elias, 2007). Within that

supply chain, there is a definite weak point during air transit. Therefore, there is a

growing interest to know which cargo is on board, where it is in the cargo hold and what

some variables such as its temperature, humidity, acceleration, etc, are. (Schmoetzer,

2005). With air freight increasing rapidly; just-in-time delivery being a real challenge;

customers and insurance providers want to know what happens to their shipment

(liability issue) and forwarders need to increase the monitoring of goods. Improvements

may be feasible for temperature sensitive goods, hazardous materials and high value

cargo, and such monitoring requires communication between the container and the

aircraft (Schmoetzer, 2005). If RFID readers are installed in the aircraft, they can

interrogate the cargo, acquire data from dedicated goods, enhance the monitoring, set

off warnings, etc. Therefore, automated onboard identification of cargo can contribute to

enhanced security (Schmoetzer, 2005).

RFID airworthiness policy. The current Federal Aviation Administration (FAA)

RFID policy document is AC20-162 "Airworthiness approval and operational allowance

of RFID systems" (FAA, 2008). It incorporates and supersedes jointly-issued AIR-100,

AFS-200 and AFS-300 RFID policy. This advisory circular offers guidance on installing

and using RFID systems on aviation products and equipment. Specifically, it provides









an acceptable way to use RFID readers or interrogators installed on aircraft, and advice

on allowing use of RFID devices on baggage, mail containers, cargo devices and galley

or service carts. It also covers using portable RFID readers or interrogators carried

onboard aircraft. The FAA airworthiness concerns about RFID systems installed on

aircraft include:

* Integrity, accuracy, and authenticity of both safety-related and identification data
from RFID devices
* Fire and electrical safety, crashworthiness, and environmental effects
* RFID device-generated RF intended transmissions or spurious emissions, both of
which can interfere with aircraft electrical and electronic systems and components
* Maintenance required for RFID devices and readers

Therefore, some of the current requirements that passive and/or active RFID must meet

are:

* Safety assessment
* Major alterations (if it might appreciably affect the aircraft's weight, balance,
structural strength, performance, flight characteristics, or other qualities affecting
airworthiness)
* Electromagnetic compatibility (EMC) demonstration (for active RFID and readers)
* Battery safety (for active RFID)
* Flammability and fire safety
* Mounting and attachment integrity
* Instructions for continued airworthiness (documentation)

This advisory circular is not mandatory and does not constitute a regulation. It describes

an acceptable (though not the only) means, to show compliance with installation and

operational requirements. Through this document, the FAA does not prohibit any use of

RFID devices, instead it provides specific requirements that the equipment must meet to

be safe and airworthy.

RF Propagation. While previous research (Rappaport and McGillem, 1989;

Valenzuela et al., 1997; Mayer et al., 2006) documents indoor propagation of radio

waves, very little work (Laniel et al., 2009) specializes on RF behavior inside a metal









environment. In contrast, the behavior of radio frequency around metal has been

studied extensively (Dobkin and Weigand, 2005; Griffin et al., 2006; Prothro et al., 2006;

Sydanheimo et al., 2006). Because aluminum is a very good conductor, incident

electromagnetic wave totally reflects from the metallic surface with a phase reversal

(Cheng, 1993; Reitz et al., 1993). Such materials are generally referred to as being "RF-

opaque". Moreover, metallic surface of the object in the vicinity of an antenna changes

its radiation pattern, input impedance, radiation efficiency and resonant frequency.

These changes depend on the size and shape of the metallic object and also on the

distance of the antenna from the object (Raumonen, 2003; Mo and Zhang, 2007). Mo

and Zhang (2007) also demonstrated that RFID tags placed 1/4 wavelength away from

the metallic surface enhances the readability of the tags. On the other hand, little or no

reflection occurs when electromagnetic waves penetrate directly through objects such

as paper, non conductive plastics or textiles (Penttila et al., 2006). These materials,

including most composites, are non-absorbing and possess low refractive indexes.

Such materials are generally referred to as being "RF-lucent".

Aircraft. Whether it is inside the cargo hold or the cabin, RFID installed inside an

aircraft would encounter a lot of metal in its environment. The entire fuselage of most

aircraft is made of aluminum alloy. Even if the use of composite materials is

continuously growing and that big steps forward have been reached on the newest

aircraft, the material distribution on an aircraft structure predominantly remains

aluminum based alloys. For example, only a mere three to four percent of the original

Airbus A300 was made of composites, but they now account for 25% of the A380

structural weight and will account for more than 50% on the future A350 (Airbus, 2009).









The shape of the metal enclosure is not only cylindrical, but also has many cross beams

as shown in Figure 4-1 which creates a very uneven reflective surface. This metal

environment leads to highly unpredictable RF propagation behavior.

DC-10-30F. According to Boeing (2010c), the multi-range DC-10 was designed

and built in Long Beach, California, by Douglas Aircraft Company, now the Long Beach

Division of Boeing Commercial Airplanes. Production was started in January 1968 and

extended to 1989, where 386 commercial DC-10s were delivered. The DC-10 Series

30F, an all-freighter model, was ordered by Federal Express in May 1984. This pure

freighter version carries palletized payloads of up to 79,380kg on more than 6,115km.

This is the model used for study in this chapter (Figure 4-2).

RFID is becoming more and more accepted for air cargo applications. Moreover,

RF propagation inside aircraft is not well documented. As a result, the objective of this

study is to evaluate the RF propagation behavior inside the cargo hold of a wide body

aircraft (DC-10-30F) at different frequencies.

Materials and Methods

Three radio frequencies (433MHz, 915MHz and 2.45GHz) were tested inside the

forward lower cargo hold of a DC-10-30 freighter aircraft (Figure 4-3). The aircraft cargo

door was kept open during the entire testing period, which simulates loading or

unloading environment. Each frequency was generated by either an RFID reader or an

RF transmitter (Table 4-1). Each RF system was connected to its respective set of

antennas as described in Table 4-1. The RF systems were installed inside the cargo

hold and the corresponding emitting antenna, connected via cable, was positioned

either at the front of the cargo hold (top end), or in the center of the ceiling (Figure 4-4).

Only one frequency and one antenna position was tested at a time. The only thing









present inside the cargo hold during testing was the RF testing equipment and one

person, who was standing on the step just outside the cargo door while operating the

spectrum analyzer. Details of the cargo hold and cargo door dimensions can be seen in

appendix A.

Test 1: Propagation Study

Signal strength data was measured in dBm (power level in decibels relative to 1

mW) via a spectrum analyzer (RSA3303B, Tektronix, Beaverton, OR) connected to the

appropriate RF receiving antenna (Table 4-1) and a 50m long LMR-400 low-loss cable.

The receiver antenna was mounted on a plastic tripod, which was moved every meter

along the length of the cargo hold at three different height and three different width

positions. This created a 3x3x12 signal strength data grid for each frequency, antenna

position and antenna type tested (Figure 4-5). The definition of a data point in this

experiment is a 200 sample "max hold" of the peak signal power observed at each

tested frequency.

Data analysis

All raw data was acquired via the spectrum analyzer in terms of signal strength

measured in dBm. Radio frequencies being tested were chosen from commercially

available products which have to obey RF spectrum regulations. In the United States,

Federal Communications Commission (FCC) regulates operating frequencies and their

respective maximum allowed output power (MAOP). Since all three systems do not use

the same output powers, antennas, connectors and cables, a calculation had to be done

to allow comparison of the systems. Link budget analysis (Shahidi, 1995; Clampitt,

2006) takes into consideration transmitter output power (Pt), measured signal strength









(Pr), transmitter antenna gain (Gt), receiver antenna gain (Gr), and various system

losses such as connectors, adapters and cables (Lsys).

RF propagation and link budget analysis

Radio signal propagation can be analyzed with various models, one of the

fundamental RF propagation equations is known as Friis transmission equation (Friss,

1946), which models line of sight propagation in free space as follow:

P = GG() (4-1)


P,: Measured signal strength
P,: Transmitter output power
G,: Transmitter antenna gain
G,: Receiver antenna gain
R: Distance of the receiver antenna to the transmitter
X: Wavelength

Wavelength is equal to the speed of light (3x108m/s) over the frequency (in Hz).

Therefore, Xo (433MHz) = 0.692m; Xo (915MHz) = 0.327m; Xo (2.45GHz) = 0.125m.

According to Friss propagation model (Eq. 4-1), Free Space Path Loss can be

calculated as (in dB, using Log base 10);

PLF,.i nep = 2OLog4- (4-2)


Equation 4-2 shows that received power is equal to the power flow through the effective

area of the receiving antenna which is also related to the wavelength and the distance

(Friis, 1946). This model (Eq. 4-2) does not account for reflections that are caused by

the high metallic environment inside the cargo hold. A better path loss (PL) model

(Nikitin and Rao, 2006) would be;









PL'= (~d) i+; -J0-e (4-3)


d: length of the direct ray path
F : reflection coefficient of the nth reflecting object
d,: length of the nth reflected ray path
N: total number of reflections

It can be seen from equation 4-3 that reflections have an important impact on path loss

calculations. The effects of reflection inside a metal environment was not added to the

link budget calculation but was instead being considered as attenuation. So far the

propagation model can be improved as follows:

P, = P,+ G, + G, PL L,, (dB) (4-4)


Where L y are the system losses due to connectors, adapters and cables. Those losses

were measured in the lab using an RF signal generator and spectrum analyzer. In this

research attenuation is more specifically defined as path loss and other losses that are

not taken into consideration such as; loss due to pointing error, atmospheric loss,

polarization loss and path loss.

Arrenuation = P~ P, + G + G, L = P-- P. (4-5)


Psys is the system's total output power. Signal strength data was put in equation 4-5 to

calculate the attenuation. Values of each parameter and calculated attenuation (Eq. 4-5)

are shown in Table 4-2. These new attenuation equations were then used to compare

results between each test. Data obtained generated a 3-D map of signal attenuation.

Full observation of the data can be seen in 12 vertical slices, as shown in the result

section in Figures 4-7 to 4-14 or in more details in Appendix B.









Link budget analysis also allows comparing data in terms of signal strength. This is

useful to evaluate the relationship between signal strength and tag reads. To be able to

compare each system to one another, it is optimal to offset the data to simulate the

maximum allowed output power (MAOP). In the case of 915MHz and 2.45GHz, the

systems used were commercially available readers that should follow FCC regulations,

but lab testing still showed deviations. On the other hand, for 433MHz, the RF

transmitter used did not have the capability to be set to the MAOP. FCC regulations are

written for measured power at a specific range. Regulation for operation in the band

433.5-434.5 MHz states (FCC, 2008):

"The field strength of any emissions radiated within the specified frequency band
shall not exceed 11,000 microvolts per meter measured at a distance of 3 meters."

Regulation for operation in the bands 902-928 MHz and 2400-2483.5 MHz states:

"For frequency hopping systems operating in the 2400-2483.5 MHz band
employing at least 75 non-overlapping hopping channels: 1 watt. For frequency
hopping systems operating in the 902-928 MHz band: 1 watt (for systems
employing at least 50 hopping channels). The conducted output power limit
specified in this section is based on the use of antennas with directional gains that
do not exceed 6 dBi."

The following calculation provides the MAOP for each frequency based on the above

FCC regulations.

P PtGt E=
PD r (4-6)


In equation (4-6), PD is the power density, E is the electric field strength and p,


(impedance of free space) = 377f. The equation can be re-written as:


Pt Gt = = 0.3 E2 (at r = 3m) (4-7)









The attenuation loss (Al) comes from logo the received power (Pr) equation:

Pr = Pt Gt Gr Al where Al = (4-8)

Therefore for 433MHz, since the maximum allowed output power measured at 3m is

11mV/m (using log base 10),

Pt Gt (at 3m) = 0.3 (0.011) = 36.3 10-6 W = 0.0363 mW

10 log(0.0 363) = -14.4 dBm

Al = 3.35 10-4 W = 0.000335 mW
(47rr)2

10log(0.000335) = -34.8 dBm

MAOP (at reader) = -14.4 (-34.8) = 20.4 dBm

For 915MHz and 2.45GHz, the limit for 1W at the reader corresponds to 30dBm, plus

6dBi antenna gain, therefore MAOP (at reader) is 36dBm. Each system's MAOP is

shown in Table 4-2. The adjustment is the difference between the MAOP and each

system's measured output power (Psys). The adjustment was added or subtracted from

the data set to mimic an optimal system and permit data comparison.

Statistical analysis

A mixed linear model was used to test the effect of frequency on attenuation. The

main effect of location of the antenna (two antenna positions) was tested on attenuation

levels for each frequency and antenna polarization; as well as the effect of antenna

polarization (circular vs. linear) for each frequency and antenna location with a mixed

analysis of variance (ANOVA) model (Littell and Milliken, 2006). The effects of width

(three vertical slices), height (three horizontal slices), and depth (12 widthwise slices) of

the receiving antenna was also tested on the attenuation level. A residual analysis was

performed to check normality and homogeneity of variance (Ott and Longnecker, 2004).









All statistical analyses were computed using SAS 9.1 (SAS Institute Inc., Cary NC) and

significance was accepted at level a = 0.01. A more conservative level of acceptance

was chosen due to the very large dataset (108 data points per test).

Test 2: Validation of Relation between Signal Strength and Tag Reads

In an ideal scenario, the measured power levels inside the cargo hold would directly

indicate the probability of having successful tag-reader communication at a specific

point. However, the cargo hold is far from being an ideal space in terms of wave

propagation due to the fact that there is interference from metals inside the cargo hold

as well as outside sources. Hence, another test was conducted as a proof-of-concept

to show that the measured power levels correspond adequately with a real RFID system

performance in terms of tag read rates.

Signal strength data acquired in Test 1 were used to compare tag readability at

915MHz using the "top end" position for circularly and linearly polarized antennas. Tag

readability was tested using 29 AD-210 Gen 1 Class 1 tags (Avery Dennison, Flowery

Branch, GA) and the ALR-9780 Genl 915MHz reader (Alien Technology, Morgan Hill,

CA). All tags were attached on a sheet of Tyvek material which covered half of the

cargo hold cross section as shown in Figure 4-6. The top of the sheet was taped to a

cardboard tube which was held by a plastic tripod. The tripod's front leg was set longer

to allow the sheet to stand as vertical as possible. This set-up was moved every meter

along the length of the cargo hold on one side (starboard or port), then it was pivoted

(along the central axis of the aircraft) to the other side to gather the other half of the

readability data. Data were acquired with the Alien software developer's kit (Alien

Technology, Morgan Hill, CA) installed on a laptop computer. Read rates obtained

correspond to the number of reads per 30 seconds.









Data point comparison

Since the signal strength readings only provide 9 data points per cross section and

the readability test uses 58 tags, the data obtained is averaged for each of the 12 slices.

Figure 4-6 shows the location of the 29 tags on the Tyvek sheet, which covers half of

the cargo hold cross section. To obtain data on the other half of the cargo hold, the

sheet pivots along the edge that is in the center of the compartment (Figure 4-6).

Results and Discussion

Test 1: Propagation Study

All raw data obtained inside the cargo hold (signal level in dBm) were offset by the

amount calculated with the link budget as shown in Table 4-2 to permit comparison in

terms of attenuation or signal strength. Comparison of raw measured signal was not

appropriate since initial output power was different from one system to the other.

Attenuation

Attenuation levels for each frequency show how the RF signal fades or attenuates

with distance. The attenuation levels observed from this experiment are shown in

Figures 4-7 to 4-14. These graphs are color coded following the visible color spectrum,

where red represents high attenuation (70dBm) and purple stands for low attenuation

(30dBm). Detailed illustration of each slice is shown in Appendix B.

Comparison of all graphs makes it apparent that 433MHz suffers less attenuation

than 915MHz, which also suffers less attenuation than 2.45MHz. As shown earlier in

this chapter, lower frequencies have longer wavelengths, which tend to travel longer

distances easier or suffer less path loss. From the graphs it can also be observed that

signal variation is more present at lower frequencies than at higher ones. Standard









deviations for each test are shown in Table 4-3. Despite the higher attenuation, a more

uniform dataset is observed at higher frequencies.

Effect of frequency on attenuation

Figure 4-15 shows a simple graph of the distribution of attenuation levels between

frequencies for circular antenna tests only (and two antenna locations). Linear antenna

testing were omitted from the comparison since, because of time constraint, there was

no such test done at 433MHz. Figure 4-15 shows a proportional relationship between

frequency and attenuation. In other words, lower frequencies lead to lower attenuation

levels inside the cargo hold. Significant difference between frequencies was statistically

tested with a mixed linear ANOVA model (p < 0.0001) (Appendix C, Table C-1).

Effect of antenna location and polarization on attenuation

The effect of antenna location as well as polarization on attenuation levels for all

frequencies mixed up was tested. The results lead to highly significant differences with

p-values <0.0001 (Appendix C, Table C-1). However, what is mostly important in reality

is the significance of this effect for each frequency and location or polarization. This is

explained in the following paragraphs.

Location. The main effect of location of the antenna (top end vs. ceiling) was

tested on attenuation levels for each frequency and antenna polarization with a mixed

linear model (Littell and Milliken, 2006). Table 4-4 shows statistical results for each test.

None of which is significantly different at a = 0.01 level.

Polarization. The effect of antenna polarization (circular vs. linear) on attenuation

for each frequency and antenna location was tested with a mixed linear model (Littell

and Milliken, 2006). Again, comparison of antenna polarization was not possible at









433MHz due to the lack of time to perform linear antenna testing on site. Statistical

analysis, as shown in Table 4-5, demonstrates a significant difference between circular

and linear polarization at 2.45GHz only. Although statistically different, the mean value

of each test differs by 2.59dBm, which may or may not have a significant effect on an

RFID system. This only depends on how close the signal level is to the sensitivity

threshold of the tags.

Effect of width, height and depth on attenuation

The effect of width, height and depth on attenuation for each frequency, antenna

location and antenna polarization was tested with a mixed linear model (Littell and

Milliken, 2006).

Width. The effect of width was only significant at 2.45GHz for the top end antenna

position and linear polarization (Appendix C, Table C-2). Statistical means were as

follows: 62.60dBm (port), 58.39dBm (center) and 61.75dBm (starboard); with p-value

<0.0001. Difference was significant between center and port as well as center and

starboard. Port and starboard were statistically similar.

Height. The effect of height was only significant at 915MHz for the top end

antenna position and linear polarization (Appendix C, Table C-3). Statistical means

were as follows: 52.66dBm (high), 46.37dBm (middle) and 45.03dBm (low); with p-value

<0.0001. Difference was significant between high and middle as well as high and low.

Middle and low were statistically similar.

Depth. The effect of depth was significant in all cases except at 2.45GHz for the

top end antenna position and linear polarization (Appendix C, Table C-4). It is

interesting to see how antenna polarization affects wave propagation. Just like it is the









case for free space propagation, linear antennas have a longer, but narrower footprint

(Dobkin, 2008). This can be observed here by the significant effect of width and height,

combined with a non significant effect of depth for linear antenna tests only. In other

words, RF signal from a linear antenna attenuates much faster as the receiver moves

sideways from the antenna, than it does as the receiver moves away in front of the

antenna. Although the width and height results were not significant for all linear antenna

tests, p-values for linear polarization was lower than for circular polarization in all cases

(Appendix C, Tables C-2 and C-3).

Signal strength

Signal strength comparison allows studying how an actual RFID system would

behave if it was installed inside the cargo hold, which will be explained in better detail in

test 2 later in this chapter. For this part of the test, the goal is to compare pure signal

propagation for each test performed (three frequencies, two antenna positions and one

or two antenna polarizations). As mentioned previously, the signal propagation data

were offset by the amount calculated in the last column of Table 4-2 (adjustment), which

brings the dataset to the maximum allowed output power as per FCC regulations. It is

important to repeat that regulations for 433MHz are much more strict that those for

915MHz and 2.45GHz. Signal propagation for each test follows the same distribution as

for attenuation since both datasets come from the same original data. However, results

show a significantly higher signal level at 915MHz compared to the two other

frequencies, with averages between -10dBm and -12dBm for 915MHz, compared to

between -21dBm and -25dBm for the other two (Table 4-6). This is due to the fact that

915MHz and 2.45GHz start with higher power levels than 433MHz; however, as

observed earlier, 915MHz has much lower attenuation levels than 2.45GHz. Therefore,









when following FCC regulations, more energy is available in the cargo hold at 915MHz.

Graphical results of each slice can be seen in Appendix D.

Test 2: Validation of Relation between Signal Strength and Tag Reads

As previously mentioned, the purpose of this test is to show that there exists a

relationship between the measured power levels by the spectrum analyzer and the tag

read rate by a commercial reader. In an ideal scenario, the power levels would be

directly proportional to read rates except irregularities such as saturation at very high

power levels. However, the substantial existence of metals inside the cargo hold will

result in antenna detuning, negatively affecting the read rates. In order to smooth out

such effect, both the read rates and the power levels across the 58 tag reads and 9

signal data points are averaged for each of the 12 cross sectional planes along the

cargo hold. This will help understand the relationship between the signal strength and

read rates while approximating the proportionality between the two quantities.

Two different antennas; circular and linear, were used for the experiment with the

same reader. The read rates and measured power levels across the 9 data points were

averaged for each of the 12 cross sectional planes. In addition, for better comparison

between the power levels and the read rates, each cross sectional power level and read

rate average are normalized by the corresponding global averages. For instance, the

global average for power levels in this study including both circular and linear antennas

is -13.5dB whereas the global average for read rates across all dimensions and antenna

types is 10.9. Hence, as Table 4-7 indicates, all the recorded values will be adjusted by

the aforementioned averages for improved statistical representation. The first column in

table 4-7 is explained as follows:

Pc: Average power levels for each cross sectional plane for the circular antenna in dB


100










Pc-adj.: Adjusted average power levels for each cross sectional plane for the circular
antenna in dB (by -13.5dB)

Rc: Average read rates for each cross sectional plane for the circular antenna

Rc-adj.: Adjusted average read rates for each cross sectional plane for the circular
antenna (by 10.9)

PI: Average power levels for each cross sectional plane for the linear antenna in dB

PI-adj.: Adjusted average power levels for each cross sectional plane for the linear
antenna in dB (by -13.5dB)

RI: Average read rates for each cross sectional plane for the linear antenna

RI-adj.: Adjusted average read rates for each cross sectional plane for the linear
antenna (by 10.9)

Table 4-7 shows that for the cross sections further from the emitting antenna, both

the power level and the read rate averages decrease in general confirming the initial

assumption that the two values are directly proportional to some extent. Furthermore,

figure 4-16 shows the adjusted average power levels and read rates for both circular

and linear antennas. In this figure, blue diamonds and red squares show the average

power levels and read rates for circular and linear antennas respectively across 12

cross sectional planes. Solid lines show the best fitted curves via linear regression with

the corresponding R-squared values. One can observe from this figure that there are

points in the system where higher average power levels do not necessarily correspond

to higher average read rates. Heavy concentration of metals around these points and

the detuning properties of the metal could have caused such discrepancies as well as

the relatively high number of tags being interrogated by the reader resulting in trafficking

problems. Nonetheless, the general trends for either antenna as well as the R-squared

values clearly show the direct proportionality between the two quantities.


101









Another useful observation is to look at the correlation coefficient between the

adjusted power level and read rate curves for both circular and linear antennas. In the

case of the circular antenna, Pcircuiar = 0.91, whereas for the linear antenna linear = 0.96.

Although both figures show a strong relationship between average power levels and

read rates for either antenna, the relationship is stronger for the linear case. This is

expected and can be explained by the fact that all the tags were carefully placed on the

Tyvek sheet in the best possible orientation with respect to the linear emitting antenna,

which confirms the fact that linear antennas perform better in use case scenarios where

tag orientation is known or controllable.

The results for both tests show that, even though there are many factors to be

considered when estimating a commercial RFID reader performance in a given

environment, measuring the signal strength at key locations in the application space

would indicate the weak and strong points in the system. It is well-known that each

system built by different manufacturers will have technical differences in terms of

sensitivity, coding technology, etc. However, this experiment indicates that the power

measurements explained in this chapter which include the three frequencies of

433MHz, 915MHz and 2.4GHz can all serve as a guideline when determining which

system would perform better in terms of RFID tag read rates under the same

circumstances and technical specifications.

Conclusion

This test demonstrated that frequencies have a major influence on signal

propagation, especially inside a metal environment. Lower frequencies suffer less

attenuation over distance, but have higher variation within the cargo hold. It was also

shown that antenna location did not deliver significantly different results. However,


102









antenna polarization can have a significant effect on signal propagation in some cases,

and therefore should not be omitted when designing an RFID system for air cargo

transportation. Moreover, FCC regulations restrict output powers at 433MHz more than

at 915MHz and 2.45GHz, which leads to the conclusion that more RF signal is available

in the cargo hold at 915MHz. It was also demonstrated that the relationship between

signal strength and tag reads is an important factor to take into account when

considering the installation of an RFID system inside an application space with nonzero

interference.


103









Table 4-1. Specifications of the three RF systems used.


Frequency


RF system


Chipcon
CC1100 RF
transmitter
433MHz (Texas
Instruments
Inc.,
Dallas, TX)


915MHz


Antenna


Type
Emitter


Polarization
Circular


Linear
Receiver (om
(omni)


915 MHz Emitter Circular
Alien RFID
reader ALR-
9780 (Alien Emitter Linear
Technology,
Morgan Hill, Linear
CA); Receiver (
(omni)


2.45 GHz
Alien RFID
reader ALB-
2.45GHz 2484 (Alien
Technology,
Morgan Hill,
CA)


Emitter Circular

Emitter Linear

Linear
Receiver (n
(omni)


Table 4-2. Calculated parameters for the attenuation equation (Eq. 4-5) and maximum
allowed output power adjustment.
Systems Pt Gt G, Lsys Attenuation MAOP Adjustement
Psys Pr for MAOP
(dBm) (dBi) (dBi) (dBm) (dBm) (dBm) (dBm)
433MHz circular 10.34 9.0 0.0 2.50 16.85 Pr 20.4 +3.55
915MHz circular 34.97 6.0 2.5 3.95 39.52 Pr 36 3.52
915MHz linear 34.97 5.9 2.5 3.95 39.42 Pr 36 -3.42
2.45GHz circular 32.24 6.0 8.0 7.26 38.99 Pr 36 -2.99
2.45GHz linear 32.24 15.0 8.0 6.71 48.54 Pr 36 -12.54

Table 4-3. Averages and standard deviations of attenuation levels for each test.
433MHz 915MHz 2.45GHz
circular circular linear circular linear
top end ceiling top end ceiling top end top end ceiling top end
Average 44.20 44.17 48.00 46.91 48.02 58.32 58.79 60.91
Std dev. 4.75 4.72 3.53 3.39 5.72 3.19 3.06 4.16


104


Gain Model & Manufacturer
9di SPA 430, Huber +
9dBi
Suhner AG, Essex, VT
B-368-1, How Tsen
Intl. Electronics Metal
OdBi Co.,Ltd. Shin Wu
Hsiang, Tao Yuan
Hsien, Taiwan
ALR-9610-BC, Alien
6dBi Technology, Morgan
Hill, CA
ARL-9610-AL, Alien
5.9dBi Technology, Morgan
Hill, CA
EXR902TN, Laird
2.5dBi Technologies,
Schaumburg, IL
2AC-001, Alien
6dBi Technology, Morgan
Hill, CA
15dBi Com-24015p, Antenna
World, Miami, FL
MRN-24008SM3,
8dBi Antenna World,
Miami, FL









Table 4-4. Statistical analysis results for the effect of antenna location.
Frequency Polarization Location Mean F value p value

433MHz Circular Top End 4170 0.00 0.9615
Ceiling 41.67
915MHz Circular Top End 4800 5.27 0.0227
Ceiling 46.91
2.45GHz Circular Top End 5832 1.20 0.2744
Ceiling 58.79

Table 4-5. Statistical analysis results for the effect of antenna polarization.
Frequency Location Polarization Mean F value p value
Circular N/A
433MHz Top End Linar N/A N/A N/A
Linear N/A
Circular 48.00
915MHz Top End circular 48.00 0.00 0.9683
Linear 48.02
Circular 58.32
2.45GHz Top End Linear 60.91 26.34 < 0.0001
Linear 60.91

Table 4-6. Signal strength data for each test, averaged per vertical slice, and total
cargo hold (Avg).
433MHz 915MHz 2.45GHz
Slices circular circular linear circular linear
(m) top end ceiling top end ceiling top end top end ceiling top end
1 -16.2 -26.6 -7.6 -14.8 -6.2 -18.6 -24.0 -24.0
2 -18.6 -23.7 -9.5 -13.5 -8.5 -19.4 -23.6 -24.7
3 E -17.3 -20.1 -10.0 -10.8 -8.9 -20.8 -22.6 -24.8
4 E -19.3 -19.6 -10.5 -11.3 -9.0 -22.4 -21.6 -23.8
5 o -18.8 -20.4 -11.2 -8.7 -9.4 -21.6 -20.3 -24.7
6 b -21.9 -16.5 -11.7 -7.3 -11.3 -21.8 -18.8 -25.1
7 w -23.1 -18.6 -11.9 -6.4 -13.1 -22.3 -19.9 -23.1
8 w -22.9 -19.7 -12.7 -9.2 -13.6 -23.4 -22.4 -25.4
9 E -24.0 -22.6 -12.3 -10.9 -16.0 -23.1 -23.2 -24.6
10 > -23.1 -22.9 -15.0 -11.4 -15.6 -24.0 -25.2 -25.6
11 -23.8 -22.2 -16.7 -12.6 -16.5 -24.7 -25.8 -26.9
12 -26.7 -22.5 -14.9 -14.3 -15.9 -25.9 -26.1 -26.2
avg -21.3 -21.3 -12.0 -10.9 -12.0 -22.3 -22.8 -24.9


105











Table 4-7. Table summarizing the recorded and adjusted power levels (P) and read
rates (R) for circular and linear antennas across the 12 cross sectional
planes.
Sectional planes
1 2 3 4 5 6 7 8 9 10 11 12
Pc -10.6 -12.5 -13.0 -13.5 -14.2 -14.7 -14.9 -15.7 -15.3 -18.0 -19.7 -17.9
Pc-adj. 2.9 1.0 0.5 0.0 -0.7 -1.2 -1.4 -2.2 -1.8 -4.5 -6.2 -4.4
Rc 18.6 14.7 16.2 14.3 13.2 9.6 4.6 5.1 1.3 0.7 0.4 0.0
Rc-adj. 7.7 3.8 5.3 3.4 2.3 -1.3 -6.3 -5.8 -9.6 -10.2 -10.5 -10.9
PI -6.2 -8.5 -8.9 -9.0 -9.4 -11.3 -13.1 -13.6 -16.0 -15.6 -16.5 -15.9
PI-adj. 7.3 5.0 4.6 4.5 4.1 2.2 0.4 -0.1 -2.5 -2.1 -3.0 -2.4
RI 20.1 19.8 20.9 16.0 18.4 13.8 11.7 11.0 8.4 7.0 6.1 10.0
RI-adj 9.2 8.9 10.0 5.1 7.5 2.9 0.8 0.1 -2.5 -3.9 -4.8 -0.9


Figure 4-1. Section of an aircraft fuselage (Airbus A380)


106




































Figure 4-2. DC-10-30F from Arrow Cargo.


PARTIAL OPEN
a-BSo


IBGROUNDLINEI


Figure 4-3. Typical fuselage section of a DC-10-30F, lower cargo hold circled in blue
(Boeing, 2010c)


107










Top End antenna position


Center Ceiling antenna position -.,


rFront


12.86m


3.18m


Back


Figure 4-4. Cargo hold dimensions and RF emitting antenna positions.


Front


Figure 4-5. Data point positions in the 3x3 grid. Twelve 3x3 grids are measured long
the length of the cargo hold, every meter.


108


7









































Figure 4-6. Tag readability test configuration. Tyvek sheet with 29 RFID tags (circled)
covering half of the cargo hold cross section.


S 30 3s M 40 45 50 [= 55 = 60 65 70


11 10


Figure 4-7. Attenuation surface plots for each vertical slice of 3x3 data point at 433MHz,
circular antenna and top end antenna position.


109


lr.
n










30 35 40 45 50 55 60 6 70


11 10


Figure 4-8. Attenuation surface plots for each vertical slice of 3x3 data
circular antenna and center ceiling antenna position.


point at 433MHz,


I 30 35s 40 45M 50 l 55s [ 60 M65 M701


12 11 10 9 8


Figure 4-9. Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz,
circular antenna and top end antenna position.


IM 30o 3s 40 45 M 50 E= 55 r- 60 8 65 70


Figure 4-10. Attenuation surface plots for each vertical slice of 3x3 data point at
915MHz, linear antenna and top end antenna position.


110


L4
7


V









Ill 30s 35s l 40 45s 5 so s5 5 60 65 701


1


11 10


Figure 4-11. Attenuation surface plots for each vertical slice of 3x3 data point at
915MHz, circular antenna and center ceiling antenna position.


S3 35 o 4 45 50 1 55 s10 65 70


11 10


Figure 4-12. Attenuation surface plots for each vertical slice of 3x3 data point at
2.45GHz, circular antenna and top end antenna position.


30 35 4 45 1 50 55 60 65 701


12 11 10


9 8


Figure 4-13. Attenuation surface plots for each vertical slice of 3x3 data point at
2.45GHz, linear antenna and top end antenna position.


111


S4 3


y










30 3 5 40 M 45 50 [ 55 860 M 65 70


11 10


Figure 4-14. Attenuation surface plots for each vertical slice of 3x3 data point at
2.45GHz, circular antenna and center ceiling antenna position.


112










20
N
=
n 15


S10


0




S10





._
M






0
0
20









N
u 15
c 10


0



I
S15
CVi
10




0


.1!,






CD LO CD LOO CD k CD UD C
C oC IT U U? W C r -_
A tt LO CDt n e n d D mCD t
Au m I I U WdBm)
Attenuadon level (dBm)


"Top End


Ceiling


Figure 4-15. Distribution (in percentage) of each frequency tested, for circular antenna
only and two antenna locations.


113


I-U1










15.00


10.00


5.00


0.00


-5 ':1:


-I0.00


15.00


* circular
* linear
-trend (circular)
-trend (linear)


0 2 4 6 8 10 12


Figure 4-16. Comparison of the change in average power levels and tag read rates for
both antennas through linear regression.


114









CHAPTER 5
RADIO FREQUENCY INTERACTIONS WITH AIR CARGO CONTAINER MATERIALS
FOR REAL-TIME MONITORING

Introduction

Products, such as food, pharmaceuticals and flowers, are at high risk of perishing

from various adversities along the cold chain. The parties involved should control when

possible, and at the very least monitor the conditions of the goods in order to ensure

their quality and to comply with all legal requirements. Among environmental

parameters during transport, temperature is the most important in maintaining the shelf

life of the products (Nunes et al., 2006; Zweig, 2006; Jedermann et al., 2009).

Cold chain. With today's globalization, there is a growing need for fresh products

to be delivered year round all over the world, thus, temperature sensitive items are likely

to be shipped by air because of their relatively short shelf life. Unfortunately, a faster

transit time does not always imply controlled temperature throughout transportation. In

contrast, during airport operations, loading, unloading, air transportation or warehouse

storage, perishable goods often suffer from temperature abuse either due to difficulties

in controlling the temperature, absence of refrigerated facilities, or lack of information

about produce characteristics and needs (Nunes et al., 2003). Of approximately 2.6

million tons of perishables air freighted in 2008, nearly 30% is estimated to be lost due

to handling and temperature abuse (Catto-Smith, 2006). In a previous study, Emond et

al. (1999) showed that the environmental conditions during airport operations could, in

fact, be very far from the optimum for fruits and vegetables. Moreover, in a strawberry

quality study, Nunes et al. (2003) showed that greater losses in quality occurred during

simulation of the airport handling operations, in-flight, and retail display than during


115









warehouse storage at the grower, truck transportation to or from the airport, or during

backroom storage at the supermarket.

Temperature monitoring. Currently, most digital temperature loggers have to be

connected to a host device to download data, and as a result, have limited real-time

data interactivity, which results in after-the-fact analysis for claims, loss in quality and

related issues. Radio frequency identification (RFID) temperature loggers function

wirelessly which allows for real-time information transfer. Active or semi-passive RFID

tags can support one or many sensors as well as the unique ID that RFID technology

provides by design. The RFID tag, with associated hardware and software gives the

added benefit of having the item scanned on receipt, so that if an alert (alert triggers are

programmable prior to shipping) is active, the receiver knows immediately (not after-the-

fact) that there is a potential problem with the shipment and can spend the time required

on specific shipments rather than going through random inspections (Jedermann et al.,

2007). Many studies have shown the effectiveness of RFID in monitoring product

temperature during transit (Emond, 2007; Jedermann and Lang, 2007; Jedermann et

al., 2007; Ketzenberg and Bloemhof-Ruwaard, 2009).

RFID technology. Although RFID's effectiveness has been proven for many

cases, the technology is not flawless. Certain materials, like metals and water-based

liquids, are challenging for RFID systems (Foster and Burberry, 1999; Emond, 2008)

and are generally referred to as being RF-Opaque. The behaviour of radio frequency

around metal has been studied extensively (Dobkin and Weigand, 2005; Griffin et al.,

2006; Prothro et al., 2006; Sydanheimo et al., 2006). Because aluminum is a very good

conductor (conductivity 38 MS/m), an incident electromagnetic wave totally reflects from


116









the metallic surface with a phase reversal (Cheng, 1993; Reitz et al., 1993). Moreover,

metallic surface of the object in the vicinity of an antenna changes its radiation pattern,

input impedance, radiation efficiency and resonant frequency. These changes depend

on the size and shape of the metallic object and also on the distance of the antenna

from the object (Raumonen, 2003; Mo and Zhang, 2007). Mo and Zhang (2007) also

demonstrated that RFID tags placed 1/4 wavelength away from the metallic surface

enhances the readability of the tags.

Not only metallic materials, but also dielectrics (or electrical insulators) cause

reflections. Other materials affect part of the incident energy and transmit the rest. The

exact amount of transmission and reflection is also dependant on the angle of

incidence, material thickness, and dielectric properties (Blaunstein and Christodoulou,

2007). On the other hand, little or no reflection occurs when electromagnetic waves

penetrate directly through objects such as paper, non conductive plastics or textiles

(Penttila et al., 2006). These materials, including most composites, are non-absorbing

and possess low refractive indexes. Such materials are generally referred to as being

RF-lucent.

Air Cargo. While the world is talking about climate change, the airline industry is

looking at ways to be more fuel efficient to minimize their operational costs as well as

their impact on the environment. One way to do so is to reduce the weight, and

minimizing weight without compromising the business volume is feasible by using lighter

containers, or Unit Load Devices (ULDs). Composite ULDs can save up to 25% of the

tare weight of a traditional aluminum ULD (Nordisk, 2010). For illustration: A Boeing

747-400 aircraft, equipped with 16 standard aluminum ULDs normally has a total of


117









1216kg empty container weight. Alternatively, by using ultra light composite ULDs, the

combined empty container weight total would be approximately 880kg. Furthermore,

composites containers are easier to repair and require fewer visits to a repair station

than aluminum units (Saunders, 2003). Kevlar ULDs are constantly replacing older

aluminum containers and account for approximately 39% of a major airline's ULD fleet.

Aluminum ULDs still add up to 43% of their fleet, whereas Lexan containers count for

the remaining 18%. Considering that ULDs have an approximate usable life of 10 years,

Aluminum ULDs will most likely be outnumbered by composite containers relatively

quickly.

This study focuses on the air transportation part of the cold chain. RFID is not yet

a widespread technology in the transportation industry, but its potential value makes it

worth the investigation effort. The objective of this study is to explore the possibility of

real time temperature monitoring during air cargo operations by researching the effect of

container wall materials on RF propagation. Five different ULD materials were chosen

for this study: Aluminum, Duralite, Herculite, Kevlar@ and Lexan. Due to the fact that

the RF behavior of materials depends on size, shape and thickness, all samples used

for this study were collected from an airline container maintenance facility and therefore

represent the true properties for each material. Initial hypotheses are that only

Aluminum samples will not allow RF transmission, whereas all other materials will

transmit radio waves with negligible interference.

Materials and Methods

Three radio frequencies (433MHz, 915MHz and 2.45GHz) were tested against five

different air cargo materials as described in the introduction: Aluminum, Duralite,

Herculite, Kevlar and Lexan. Duralite is a thick fibreglass woven composite. Herculite


118









(or Twintex P PP) is a thermoplastic glass reinforcement panel made of commingled

E-Glass and thermoplastic filaments. Kevlar@ is made with high strength para-aramid

fiber and Lexan is a translucent polycarbonate. For the first two tests, each sample

was a square sheet of 0.305m long sides and thicknesses of 1.00, 1.80, 1.00, 0.50 and

1.80mm respectively. For the third test both samples were squares of sides 1.22m long.

This series of tests was performed inside an anechoic chamber of dimensions

2.05m high, 1.90m wide and 2.70m deep. The wall materials were Eccosorb VHP-12-

NRL and Eccosorb FS-100-NRL (Emerson & Cuming Microwave Products N.V.,

Westerlo, Belgium), a solid, pyramidal shaped, carbon loaded urethane foam absorber.

Each frequency was generated by an RF signal generator (Agilent N9310A, Agilent

Technologies, Santa Clara, CA); power supply (XTR 33-25, Xantrex technology,

Burnaby, BC, Canada); and power amplifiers (5803039A and 5303081, Ophir RF, Los

Angeles, CA). This equipment was located outside of the anechoic chamber during

testing. The RF output of this system was conveyed to the anechoic chamber via a 50m

long LMR-400 low-loss cable. Each frequency was tested with a particular set of emitter

and receiver antennas (Table 5-1) and only one frequency was tested at a time.

Three tests were administered to determine the effects of the materials on RF

propagation. For all tests, the received signal was measured with a spectrum analyzer

(RSA3303B, Tektronix, Beaverton, OR), also kept outside the door of the anechoic

chamber during testing. The definition of a data point in this experiment is a 200 sample

average of the peak signal power observed at each tested frequency. One frequency

was tested at a time and all data were analyzed with reference to the control data point

(no material sample present).


119









Test 1

The goal of this test was to quantify the reflection and absorption characteristics of

each material. Inside the chamber, the emitting antenna, receiver antenna and material

samples were arranged in a row on a Plexiglas table with a Styrofoam plate to help hold

everything in place (Figure 5-1). The table was centered in the room 0.38m above the

floor, just over the anechoic chamber surface material responsible from absorbing

outside RF radiation. The emitting antenna was positioned vertically, beaming towards

the back of the room. The receiving antenna was also positioned vertically with specific

intervals based on the radiation wavelength (at X/2, 3X/4, k+X/4 or k+X/2), and the

material sample was positioned at X in front of the emitting antenna.

c
Wavelength in meters is calculated as; A =- (5-1)
f

Where c is the speed of light in m/s and f is the frequency in Hz. In other words, the

sample was 0.692m from the 433MHz antenna; 0.328m away from the 915MHz

antenna; or 0.125m away from the 2.45GHz antenna. The respective receiver antennas

were consecutively placed 4 and 12 wavelengths away from the sample, on both sides

(Figure 5-1). Ideally, test 1 should have been accomplished with infinite planes of each

sample. Reality is different, and material availability was limiting. This design is

interesting in the way that it procures information on more aspects of radio frequency

behavior, such as wave scattering and diffraction around sharp obstacles. In reality,

those effects exist and are inevitable components in RFID applications.

Test 2

In order to achieve a more uniform dataset, the goal of this test was to isolate the

receiver antenna from the knife-edge diffraction effect. The sample was framed with a


120









solid, pyramidal shaped carbon loaded urethane foam absorber (anechoic chamber wall

material). The foam pyramids were glued onto a 0.05m thick Styrofoam sheet and were

positioned to leave the center part of the 0.305m by 0.305m square empty to place the

samples as shown in Figure 5-2A. The samples were again positioned one wavelength

away from the emitting antennas, and the receiver antenna was taped behind the

sample (Figure 5-2B). Three repetitions of each data point were performed for statistical

analysis. Statistical analysis consisted of one-way ANOVA to show significant

differences between the materials for each frequency. Multiple comparisons of means

were performed with Bonferroni adjustments. All statistical analyses were computed

using SAS 9.1 (SAS Institute Inc., 2003).

Test 3

Following the thought process from test 1 to test 2, it was determined that an

additional test was needed to clarify the effect of using a smaller sample and therefore

show more realistic properties of RF-lucent and RF-opaque materials. Since larger

samples were not available in all materials, Aluminum and Kevlar were chosen and

samples of 1.22m x 1.22m were tested (which is 16 times larger than the previous

sample size). Those two materials were chosen because of availability, but also

because of their wide use in the air cargo container fleet. Aluminum and Kevlar

containers cover together over 80% of today's container fleet at a major airline.

Moreover, those two materials can represent typical RF-lucent and RF-opaque

materials encountered in the air cargo industry.

The same set-up as test 1 was used, which means that the emitting antenna was

positioned vertically, beaming towards the back of the room. The receiving antenna was

also positioned vertically with specific intervals based on the radiation wavelength (at


121









X/2, 3X/4, Xk+/4 or X+X/2); and the material sample was positioned at X in front of the

emitting antenna. Except this time because of sample size, the antennas and samples

were hung from the ceiling of the chamber with strings, instead of being held in place on

the table. Three repetitions of each data point were performed for statistical analysis

Results and Discussion

Test 1

The results showed a very strong effect for Aluminum on RF transmission, and

minimal interaction for all other sample materials. All comparisons were made between

the control and each sample. Table 5-2 shows values obtained for the control

measurements (no material present), whereas Table 5-3 illustrates the signal deviations

from the control (control subtracted from each signal strength measurement). Receiver

antenna positions are measured from the emitting antenna and sample materials are

positioned at X.

433MHz

Results show weaker signal levels in front of the Aluminum sample at X/2 (-

1.52dBm) and higher signal strength at 3X/4 (+2.39dBm) (Table 2-3). This confirms the

observation made by Mo and Zhang (2007), which is when an electromagnetic wave

hits a metallic surface, it reflects with a 1800 phase reversal. This causes signal

cancellation at X/2 and signal amplification at X/4 from the metallic surface. In our case,

X/4 from the Aluminum sample is 3X/4 distance from the emitting antenna. All other

samples show no considerable loss or gain from reflections when the receiver antenna

was positioned in front of the samples (within 0.09dBm from control). As far as signal

transmission through the samples, it is understandable that only the Aluminum sample


122









offers considerable signal blocking, with signal loss of -5.45dBm at kX+/4 and -2.70dBm

at X+X/2. All other materials were within 0.50dBm from the control.

915MHz

In this part of the experiment, signal strength in front of the Aluminum sample was

increased in both X/2 and 3X/4 cases, although the increase was greater at 3X/4 (+7.11

vs. +3.98). This could be caused by signal scattering since the plate size (0.305m) was

slightly smaller but very close to the wavelength at 915MHz (0.325m). Since for the

case of 915MHz the wavelength and the dimensions of the material (obstacle) were of

similar sizes, the set-up was in the resonance range (Finkenzeller, 2003). Therefore, the

behavior of RF radiation may not follow traditional rules such as the one stated by Mo

and Zhang due to the unpredictable nature of edge diffractions (Longhurst, 1967) as

well as resonance. Signal was also slightly reflected from of other materials, Lexan

being the second most reflecting with +1.18dBm gain. In the case of signal

transmission, similar results were observed as with 433MHz, except the signal loss is

greater, with -19.70 and -11.74 at X+X/4 and X+X/2 respectively. All other materials were

within 0.17dBm of the control.

2.45GHz

Results for this part of the experiment follow Mo and Zhang's theory of wave

reflection with a loss of -6.86dBm and a gain of +2.81dBm at X/2 and 3X/4 respectively.

All other materials were within 0.58dBm of the control. Moreover, the signal loss

behind the samples was obvious with -37.99dBm at kX+/4 and -34.37dBm at kX+/2, all

other materials being within 0.55dBm of the control.


123









Looking at signal transmission behind the Aluminum sample, it was noticeable that

the signal loss increases with the frequency. This was caused by the ratio of the

wavelengths and the materials sample size. At 433MHz, the wavelength was more than

two times longer than the sample size (0.692m and 0.305m respectively); at 915MHz,

both dimensions were similar (X = 0.325m); and at 2.45GHz, the wavelength was about

half of the sample size (X = 0.125m). When a radio wave impinges an obstacle larger

than its wavelength, reflection occurs. However, when a wave hits an obstacle smaller

than its wavelength, scattering occurs and wave patterns are redirected with random

phase and amplitude (Blaunstein and Christodoulou, 2007).

It is also noticeable that there was minor signal amplification behind the non

metallic samples at 433MHz and 915MHz. This can be explained by the fact that the

sample size was smaller than the wavelengths, which allows waves to travel around the

materials' edges. This effect is known as the knife-edge diffraction and explains the

redirection of electromagnetic waves when they hit a solid obstacle such as the edge of

the material sample in this experiment (Kumar et al., 2007). Knife-edge diffraction is

described by Huygens-Fresnel principle which states that such an obstruction (the edge

of the material in this case) will act as a secondary source of RF radiation (Longhurst,

1967). Depending on the wavelength of the electromagnetic signal, the effects of this

secondary source could be observed at different points in the measurement field, in this

case, amplification behind the non-metallic samples, however, the discussion of this

phenomenon in greater detail is beyond the scope of this text.


124









Test 2

When six treatments (material samples) were compared, all results are reported

as significant when P < 0.05 and the Aluminum sample was the only one significantly

different from the others for all three frequencies. Due to the nature of the second

experiment it would be expected to obtain higher attenuation at lower frequencies

because shorter wavelengths travel more easily inside the open frame within the foam

absorber material. However, one should note that this observation is affected by two

important parameters: the electromagnetic properties of the container samples as well

as the absorption profile of the urethane foam absorber, which is proportional to the

frequency (Eccosorb, 2008). For instance, for free air (control) the signal strength at

433MHz is 3.83dBm whereas the signal strength at 915MHz is 9.53dBm. This clearly

shows the attenuation from the wavelength dimension at lower frequency as expected.

However, at 2.45GHz, the signal power was attenuated to 6.96dBm, which is explained

by the fact that the foam absorber material has higher absorption coefficients at higher

frequencies.

Test 3

Table 5-5 shows values obtained for the control measurements (no material

present), whereas Table 5-6 illustrates the signal deviations from the control (control

subtracted from each signal strength measurement). All comparisons were made

between the control measurement and each sample. The results showed a very strong

effect for Aluminum on RF transmission, and minimal interaction for Kevlar as

expected from the previous tests. This test, however, showed a much more important

attenuation level behind the Aluminum sample when comparing to the results obtained

in test 1 (same test set-up, different sample sizes) for frequencies 433MHz and


125









915MHz. As stated previously, 433MHz corresponds to a wavelength of 0.692m,

whereas 915MHz corresponds to 0.328m and 2.45GHz to 0.125m. As opposed to the

earlier tests, the dimensions for the samples in test 3 were 1.22m x 1.22m, which are

larger than all three wavelengths in this case.

Similar to the previous tests, the attenuation levels measured behind the

Aluminum sample were still proportional to the frequency tested. Attenuation at 433MHz

is lower than attenuation at 915MHz and 2.45GHz, but the divergence was not as

strong as in test 1 which can be explained by the use of a larger sample between the

emitting and receiving antennas. The general trend shows that an infinite aluminum

plane would probably lead to similar results in terms of total attenuation for all

frequencies. As far as reflection, a similar pattern was observed where the RF waves

that reflects from the aluminum surface increase the signal level in front of the sample in

all cases. This confirms again what Mo and Zhang (2007) had previously observed,

which is when the electromagnetic wave reflects off a metallic surface, it causes signal

cancellation at k/2 and signal amplification at k/4 from the metallic surface.

As previously stated, the goal of this test was to show that a larger-than-

wavelength sample size would affect the signal level measurements behind the samples

by eliminating secondary effects such as edge reflection discussed in the previous

sections and uniformize the results. The values presented in tables 5-5 and 5-6 show

that this goal was accomplished by measuring higher power levels in front of the sample

material for Aluminum and behind the sample material for Kevlar.


126









Conclusion

This test demonstrated the effects of five commonly used air cargo container wall

materials on RF propagation at three different frequencies. The reflection and

absorption characteristics of each material were quantified. Three different tests were

utilized to analyze the characteristics of RF propagation in greater detail for each

material and the results from all experiments showed a very strong effect of Aluminum

on RF transmission and minimal interaction for all other sample materials as expected.

These findings suggest that the use of non-metallic containers for air transportation of

perishable products should make real time temperature monitoring possible by allowing

RF waves to transmit through the wall surface effortlessly. This goes well with the

current trend that encourages the use of Kevlar@ containers over aluminum ones

because of their much lighter tare weight.


127









Table 5-1. Specifications of the six antennas used.
Frequency Antenna Polarization Gain Model & Manufacturer
Em r C r 9 di SPA 430, Huber + Suhner AG,
Emitter Circular 9 dBi
Essex, VT
433MHz B-368-1, How Tsen Intl. Electronics
Receiver mni) 0 dBi Metal Co.,Ltd. Shin Wu Hsiang, Tao
Yuan Hsien, Taiwan
SPA 915, Huber + Suhner AG,
Emitter Circular 8 dBi
915MHz Essex, VT
9 z Linear EXR902TN, Laird Technologies,
Receiver 2.5 dBi
(omni) Schaumburg, IL
Emitter Circular 6 dBi 2AC-001, Alien Technology, Morgan
Emitter Circular 6 dBi
2.45GHz Hill, CA
5Linear MRN-24008SM3, AntennaWorld,
Receiver 8 dBi
(omni) Miami, FL

Table 5-2. Signal strength measurements (dBm) for control (no sample), test 1.
Receiver antenna positions are measured from the emitting antenna.
Frequencies Receiver antenna positions
k/2 3X/4 X+X/4 X+X/2
433MHz 11.07 9.25 8.65 1.80
915MHz 13.90 11.03 7.97 6.15
2.45GHz 8.94 7.85 6.64 6.12

Table 5-3. Signal strength deviation (dBm) from control (no sample) for test 1. Receiver
antenna positions are measured from the emitting antenna and sample
materials are positioned at X.


Frequencies Materials


Aluminum
Duralite
Herculite
Kevlar
Lexan
Aluminum
Duralite
Herculite
Kevlar
Lexan
Aluminum
Duralite
Herculite
Kevlar
Lexan


k/2
-1.52
0.02
-0.03
-0.02
0.01
3.98
1.09
0.84
0.96
1.18
-6.86
0.22
-0.17
-0.15
-0.07


Receiver antenna positions
3X/4 k+k/4
2.39 -5.45
0.09 0.02
0.02 0.15
0.02 0.09
-0.02 0.08
7.11 -19.70
0.40 0.17
0.30 0.08
0.32 0.09
0.16 0.17
2.81 -37.99
0.38 -0.32
0.54 -0.51
0.30 -0.55
0.58 -0.29


k+k/2
-2.70
0.18
0.50
0.40
0.14
-11.74
0.12
0.03
0.00
0.10
-34.37
-0.10
-0.50
-0.16
-0.23


128










Table 5-4. Signal strength measurements (meanSD) (dBm) for control, plus signal
strength deviation between material samples and control at three frequencies
for test 2 (n=3).


433MHz
3.830.05
-15.160.06
-0.050.03
-0.040.03
-0.010.01
-0.060.04


Frequencies
915MHz
9.530.05
-20.490.18
-0.210.01
-0.290.02
-0.350.01
-0.530.65


2.45GHz
6.960.01
-35.470.23
-0.190.03
-0.330.01
-0.370.01
0.070.03


Table 5-5. Signal strength measurements (dBm) for control (no sample), test 3.
Receiver antenna positions are measured from the emitting antenna.
Frequencies Receiver antenna positions
k/2 3k/4 k+k/4 k+3/2
433MHz 10.60 8.50 4.53 1.97
915MHz 14.93 11.72 7.76 6.41
2.45GHz 11.04 8.97 5.47 4.79

Table 5-6. Signal strength deviation (dBm) from control (no sample) for test 3. Receiver
antenna positions are measured from the emitting antenna and sample
materials are positioned at X.
Frequencies Materials Receiver antenna positions
k/2 3X/4 X+X/4 X+X/2
Aluminum 0.36 3.18 -20.01 -15.33
433MHz
Kevlar 0.74 0.30 -1.48 1.27
Aluminum 2.17 6.20 -26.84 -24.53
915MHz
Kevlar 0.19 -0.42 -0.13 0.11
Aluminum 0.84 3.20 -38.12 -37.48
2.45GHz lar0.230.400.000.56
Kevlar 0.23 0.40 0.00 0.56


129


Materials

Control
Aluminum
Duralite
Herculite
Kevlar
Lexan












RFID receiving antennas


Plexiglas structure I Styrofoam

Anechoic chamber
RFID Emitting Antenna ..surface materials











Figure 5-1. Diagram of the anechoic chamber setup for test 1. Note that four receiver
antennas are shown for illustrative purposes as only one receiver antenna is
used at a time for each test.


- Material Sample


Figure 5-2. Anechoic chamber set-up for test 2. A) The sample material is surrounded
by pyramidal shaped, carbon loaded urethane foam absorber and is placed
one wavelength from the emitting antenna. B) The receiver antenna is taped
behind the material sample.


130









CHAPTER 6
TEMPERATURE MAPPING INSIDE AIR CARGO CONTAINERS DURING AIRSIDE
OPERATIONS

Introduction

Temperature is well regulated in the cabin of most passenger flights, but it is not

necessarily the case inside the cargo hold or in freighter flights. Temperature

distribution and variability in cargo holds depends on many factors such as weather (air

temperature, wind speed, sun radiation), duration of flight, type of aircraft (ability to

control cargo ambient temperature), altitude, and transit time on the tarmac.

Aircraft. Temperature control inside the cargo hold can be regulated at different

levels. On one end, the only heat available in the cargo hold of some aircraft comes

from leaks in the cabin floor; some other airplanes have ventilation systems that can re-

circulate the cabin air into the cargo hold; and the more sophisticated ones have their

own heating systems designed to keep the cargo from freezing (Air Canada, 2007).

Typically, larger and newer aircraft have more options. When aircraft are categorized

by size, two major categories emerge, which are narrow-body and wide-body aircraft.

Narrow-body aircraft have one aisle and two rows of seats in the cabin, whereas wide-

body aircraft have two aisles and three rows of seats in the cabin floor. As a

consequence, the wide-body aircraft has a larger cargo compartment and can also be

used on longer routes. Studies on temperature profiles inside cargo holds state many

different temperature ranges. Syversen et al. (2008) found that 49.5% of shipments

were exposed to high temperatures (greater than 29.4 oC), 14.6% to low temperatures

(less than 7.2 oC), and 61% to temperature variations of 11 oC or more. It was also

shown that temperature depends on ULD position inside the cargo hold as well as

which cargo hold is used (Emond et al., 1999).


131









Air cargo operations. All cargo being planned on a flight is built-up onto unit load

devices (ULDs) or in tub carts (for bulk) a specific time prior to flight departure. This

specific time depends on the airport, destination (domestic vs. international), size of

freight, and type of product (priority vs. general freight) (T. Howard, 2010, personal

communication). The freight usually leaves the warehouse (on the airside) for a

maximum of 2h before flight time, but typically spends between 45 and 90 minutes on

the tarmac before being loaded onto the aircraft (T. Howard, 2010, personal

communication). In practice, if something happens to delay the loading activity of the

plane, the freight can be left on the tarmac for extended periods of time, regardless of

weather conditions (Villeneuve, 2006). Many factors can influence ULD's temperature

fluctuations while waiting on the tarmac, for example, sun radiation, air temperature,

ULD wall material properties, wind speed and direction, etc (Villeneuve et al., 2001).

Perishables. Products, such as food, pharmaceuticals or flowers, are at high risk

of perishing from various adversities along the cold chain. Among environmental

parameters during transport, temperature is the most important in maintaining the shelf

life of the products (Nunes et al., 2006; Villeneuve, 2006; Zweig, 2006; Jedermann et

al., 2009). With today's globalization, there is a growing need for fresh products to be

delivered year round all over the world, Temperature sensitive items are likely to be

shipped by air because of their relatively short shelf life. Unfortunately, a faster transit

time does not always imply controlled temperature throughout transportation. In

contrast, during airport operations, loading, unloading, air transportation or warehouse

storage, perishable goods often suffer from temperature abuse either due to difficulties

in controlling the temperature, absence of refrigerated facilities, or lack of information


132









about produce characteristics and needs (Nunes et al., 2003). On approximately 2.6

million metric tons of perishables air freighted in 2008, nearly 30% is estimated to be

lost due to handling and temperature abuse (Catto-Smith, 2006). In a previous study,

Emond et al. (1999) showed that the environmental conditions during airport operations

could, in fact, be very far from the optimum for fruits and vegetables. Moreover, in a

strawberry quality study, Nunes et al. (2003) showed that greater losses in quality

occurred during simulation of the airport handling operations, in-flight, and retail display

than during warehouse storage at the grower, truck transportation to or from the airport,

or during backroom storage at the supermarket. Many more studies showed that

important temperature fluctuations can occur during airport ground operations (Bollen et

al., 1998; Villeneuve et al., 2000; Villeneuve et al., 2001). Moreover, because

perishables are mostly season dependant, transportation companies do not offer

special treatments like they would if they were available year round (Villeneuve, 2006).

Temperature monitoring. Currently, most digital temperature loggers have to be

connected to a host device to download data, and as a result, have limited real-time

data interactivity, which result in after-the-fact analysis for claims, loss in quality and

related issues. Radio frequency identification (RFID) temperature loggers function

wirelessly which allows for real-time information transfer (Rao, 1999, Lahiri, 2006).

Active or semi-passive RFID tags can support one or many sensors as well as the

unique ID that RFID technology provides by design. The RFID tag, with associated

hardware and software will add the benefit of having the item scanned on receipt, so

that if an alert (alert triggers are programmable prior to shipping) is active, the receiver

knows immediately (not after-the-fact) that there is a potential problem with the


133









shipment and can spend the time required on specific shipments rather than going

through random inspections (Jedermann et al., 2007). Many studies have already

shown the effectiveness of RFID in monitoring product temperature during transit

(Emond, 2007; Jedermann and Lang, 2007; Jedermann et al., 2007; Ketzenberg and

Bloemhof-Ruwaard, 2009).

Since temperature control inside cargo holds is a weak link in the air cargo cold

chain, the objective of this study is to evaluate the temperature distribution inside the

ULDs during airside operations. Comparison will be made between long and short

flights, as well as between wide- and narrow-body aircraft. Hypotheses are that

temperature will drop lower at the bottom part of containers as well as during longer

flights in general.

Materials and Methods

In June 2010, 12 ULDs were shipped on 10 different one-way flights (Table 6-1).

The number of ULDs monitored per aircraft was limited by the amount of cargo

available to travel on that route that specific day. All flights were managed by Air

Canada and originated from Toronto, Canada (YYZ). Three ULDs traveled to Montreal,

Canada (YUL); one ULD flew to Vancouver, Canada (YVR); four ULDs flew to London,

UK (LHR); and four ULDs were shipped to Frankfurt, Germany (FRA). All these ULDs

were reloaded at destination to be shipped back on a returning flight. All containers

shipped on Canadian routes were shipped back to Toronto, however, ULDs shipped to

Europe either returned to Montreal, Toronto or Vancouver. All flight times varied

between 1h and 9.7h (Table 6-1).

ULD capacity. As stated in Table 6-1, domestic flights were made onboard Airbus

320 or 321 narrow-body aircraft, which hold LD3-45 ULDs (or called by their prefix


134









AKH). These ULDs take the entire width of the cargo hold. The A320 can hold three in

the forward compartment and four in the aft compartment. Being slightly longer, the

A321 can hold five ULDs in the front as well as four in the aft cargo hold. On the other

hand, international flights were made on wide-body aircraft which carry LD3 ULDs

(prefix AKE). Those ULDs occupy half of the cargo hold width and are loaded side by

side. The A330 can hold 18 in the front and 15 in the aft compartment, whereas the

Boeing 777 can carry up to 24 in the front and 20 in the aft compartment. Five tags were

placed in each AKH and eight tags were installed in each AKE container according to

the scheme shown in Figure 6-1.

Temperature control. According to Air Canada Load Control Engineering

publications, there is no ventilation and no heating in any of the cargo holds of the A320

and A321 aircraft except for cargo door leakage when there is a pressure differential

between the fuselage interior and exterior. The Airbus specification for the A321 aircraft

guarantees a minimum temperature of 20C in flight (Air Canada, 2005a, b). In the A330,

the forward cargo hold is equipped with a temperature controlled heating and cooling

system that is ventilated at all times. Under normal conditions, the mean in-flight

temperature will be between 5C and 250C. The aft cargo compartment is neither

provided with ventilation nor heating system (Air Canada, 2006). In the B777, all cargo

holds are heated. The forward hold is equipped with an air conditioning system

designed to maintain a constant target temperature and provide ventilation both on the

ground or during flight. A temperature selector in the cockpit provides a selectable

temperature control ranging from 40C to 270C. The aft hold is equipped with a basic

heating system providing compartment temperature control to two set points


135









corresponding to settings of LOW (40C to 10C) and HIGH (18C to 240C) (Air Canada,

2007).

Temperature sensors. The temperature sensors used for monitoring were

TurboTag (Sealed Air Corporation, Elmwood Park, NJ), which are high frequency

(13.56MHz) RFID tags with temperature logging capacity of 702 time-temperature data

points. Tag accuracy is 0.5C throughout normal operating range (-250C to +350C) at

95% confidence interval. They were programmed to read every five minutes, which

allowed close to 2.5 days of monitoring time. All tags were started the morning before

the first flights and stopped automatically when all 702 points were recorded. Data was

downloaded at the end of the experiment.

It is important to note that all "during flight" data shows temperature recorded

between the time the aircraft left the origin gate, and the time it sets the brakes at the

destination gate, thus including taxi, takeoff and landing. Moreover, "airside operations"

refers to everything between the time the cargo leaves the warehouse to go on the

tarmac, until it comes back to another warehouse for customer pick-up.

Results and Discussion

During Flight

On a general basis, sensors recorded much lower temperatures in the bottom of

ULDs than in the upper part during flight (Figure 6-2). This observation supports one of

the initial hypotheses. Moreover, this fact can be explained by a combination of factors

such as the distance from the aircraft skin, the heat coming from the passenger cabin

and natural convection.

The narrow-body aircraft used in this study were the Airbus 320 and 321 and are

considered as short to medium range aircraft. As explained earlier, these aircraft do


136









not have any means for heating their cargo holds, and consequently temperature

distribution depends entirely on outside/surrounding conditions. For narrow-body

aircraft, this study shows that the duration of the flight significantly affects temperature

drop (Figure 6-2). Flights under 2h (Toronto-Montreal) stayed above 15C all around

(Figure 6-3), whereas flights to and from Vancouver (4-6h) dropped to 3C at the bottom

of the ULDs (Figure 6-4). Mostly, it is the temperature on the floor of the cargo hold that

cools the most, but the overall temperature also drops. Cold or warm temperature does

not necessarily mean good or bad. Temperature sensitive goods do not always require

refrigeration. Tropical fruits, for example, should never be exposed to close to freezing

temperatures, and berries, on the other hand, should be kept as close to 1C as

possible.

For wide-body aircraft (A330 and B777), temperature shows a decreasing trend

while in the air (Figure 6-5 and 6-6). When averaging all curves on Figure 6-5, the

general slope becomes negative after only 1h. As a result, flight time is inversely

proportional to overall temperature in the cargo compartments, at least for flights over

7h. Nonetheless, the temperature drop during flight was steeper for the narrow-body

(Figure 6-4) for the same length of time when comparing with the first 5h of wide-body

aircraft' flights (Figure 6-5, 6-6). This could be due to the temperature control ability of

wide-body aircraft. However, the only ULD carried in the narrow-body A321 (Figure 6-

4) was placed at the same position on both flights (Figure E-2), therefore results could

change with data from other locations inside the aircraft.

For shipments in the A330, some ULDs were carried aboard the forward or aft

compartment, which means that some loads were "heated and ventilated" while some


137









were not. Averaging all tag temperatures carried in the forward (heated) compartment

during flight compared to those transported in the aft (unheated) cargo hold (Table 6-2)

leads to the conclusion that there is no significant difference in temperature profiles.

This conclusion might be biased by the fact that all ULDs transported in the heated

compartment were positioned at the very front of the aircraft, right next to the cargo

door. Moreover, if there was no requirement for heating the cargo hold during those

specific flights, the system would have been set to default, which is ventilation "on" but

no heat. Those two factors could have strongly contributed to this almost homogeneous

result, which is counterintuitive.

Before and After Flight

ULDs are brought to the gate up to 2h before each flight, and can typically spend

the same amount of time waiting on the tarmac after being unloaded from the aircraft

(an ideal, no delay situation) (T. Howard, 2010, personal communication). But as the

present study showed, even a short period of time waiting on the tarmac can lead to

very high temperatures on top of the UDLs (in summer season). Maximum recorded

temperatures at the top of the ULD peaked at 45C in less than 20 minutes from arrival

at the gate (see 7h-mark in Figure 6-7). The goods pulp temperature certainly do not

warm up as fast, but sensitive shipments placed near the top of the containers could

easily suffer from a major break in the cold chain. On the other hand, goods loaded in

the bottom part of the container would take much longer to experience the temperature

raise. Other examples can be found in appendix E where each ULD's temperature

profile (2h before first flight until 2h after second flight) is plotted separately. Each graph

also shows the container positions inside the aircraft. Throughout the year, the effect of


138









ground weather on ULD temperature should see much more variation than during the

in-flight transportation since air temperatures at high altitudes do not vary as much.

Conclusion

Cold chain is always hard to keep intact when goods transfer from one hand to

another, especially when the transportation methods do not offer refrigeration during

transit. This temperature distribution study showed that a major temperature gradient

can be found within the same ULD during tarmac operations as well as during flight,

especially when the flight time exceeds 4h. Moreover, temperature seems to drop faster

inside the cargo holds of narrow-body aircraft, but further similar studies are required

to verify that statement. Since this study was performed in the summer, it would be both

interesting and informative to measure temperature distribution and variability for similar

tests performed during the winter. Furthermore, one could investigate transportation

with longer flights (>10h) to measure temperature variability and distribution and to see

if temperature in the bottom of the ULDs would reach critical freezing temperatures.


139










Table 6-1. Routes, aircraft and ULD specs from Toronto (YYZ).
ULDs Aircrafts Flight time
Destination 1 Destination 2 Us A F
qty type to/from to / from (h)
YUL YYZ 3 AKH Airbus 320 1.1/1.4
YVR YYZ 1 AKH Airbus 321 5.4/4.3
LHR YUL 2 AKE Airbus 330 7.2 / 7.5
LHR YYZ 1 AKE Airbus 330 7.2 / 8.0
LHR YVR 1 AKE A330/B777 7.2 / 9.7
FRA YUL 4 AKE Boeing 777 7.5 /7.5

Table 6-2. Temperature comparison between heated and unheated cargo holds inside
an Airbus 330.
Heated compartment Unheated compartment
Flight AKE# Temperature Flight AKE# Temperature
Flight AKE # means (C) Flight AKE # means (C)
means (oC) means (oC)
Top bottom Top bottom
tags tags tags tags
YYZ-LHR 03782 19.20 12.71 YYZ- LHR 04090 17.60 13.97
YYZ-LHR 04969 21.26 15.58 LHR-YYZ 04090 20.20 14.71
LHR-YUL 04969 20.78 13.18 YYZ- LHR 05335 17.54 14.01
LHR-YUL 05335 22.02 16.79
Total averages 20.41 13.82 Total averages 19.34 14.87
17.12 17.11


AKH or LD3-45 (5 tags)


Figure 6-1. ULD types and their respective tag positions.


140













25


20 -




0-10
E
0--


Top


5 -
1-2h 4-6h 7-8h >9h
Narrow-body Wide-body

Total flight time (h)

Figure 6-2. Temperatures recorded for top and bottom tags during flight (gate to gate).
Data is congregated by total flight time and type of aircraft. The boundary of
the box closest to zero indicates the 25th percentile, a line within the box
marks the median, and the boundary of the box farthest from zero indicates
the 75th percentile. Error bars above and below the box indicate the 90th and
10th percentiles.


141














20

-. -TopYUL-YYZ

S15 -BottomYUL-
YYZ
-TopYYZ-YUL
E 10
E0 Botom YYZ-
YUL

5


0
0:05 0:15 0:25 0:35 0:45 0:55 1:05 1:15 1:25
Flighttime (h)


Figure 6-3. Graph of averaged top and bottom tag temperatures during flight (gate to
gate) for both short flights (1-2h) to and from Montreal (YUL).

25


20

-Top YYZ-YVR

15 -BottomYYZ-
YVR
\ -TopYVR-YYZ
E 10
S10 BottomYVR-
YYZ





0
0:05 0:35 1:05 1:35 2:05 2:35 3:05 3:35 4:05 4:35 5:05
Flighttime (h)


Figure 6-4. Graph of averaged top and bottom tag temperatures during flight (gate to
gate) for both medium-short flights (4-6h) to and from Vancouver (YVR).


142











-TopYYZ-FRA
-BottomYYZ-
FRA
20 -Top FRA-YUL

Bottom FRA-
15 l-TopYYZ-LHR
| -BottomYYZ-
LHR
E 10 -Top LHR-YYZ
SBottom LHR-
YYZ
5 -Top LHR-YUL
-Bottom LHR-
YUL
0
0:05 0:55 1:45 2:35 3:25 4:15 5:05 5:55 6:45 7:35
Flighttime (h)

Figure 6-5. Graph of averaged top and bottom tag temperatures during flight (gate to
gate) for all 7-8h flights to and from London (LHR) or Frankfurt (FRA). Red-
pink colors are for top temperatures, and blue-green colors are for bottom
temperatures.

30


25


C-20
--Top LHR-YVR
S15
Bottom LHR-
E YVR
S10


5


0
0:05 1:05 2:05 3:05 4:05 5:05 6:05 7:05 8:05 9:05 10:05
Flighttime (h)

Figure 6-6. Graph of averaged top and bottom tag temperatures during flight (gate to
gate) for the longest flight (above 9h) between London (LHR) and Vancouver
(YVR).


143









YUL-YYZ


50
45
40
0 35
S30
30
c 25
. 20
E
w 15
10
5
0


LZ



L


? I


Time (h)
Time (h)


Figure 6-7. Temperature profiles of all tags for ULDs AKH 1817 to and from Montreal.
First flight segment is highlighted in yellow and returning flight segment is
highlighted in blue. Corresponding container positions are shown on the
sketch to the right.


144


AKH 1817


YYZ-YUL










CHAPTER 7
GLOBAL TRACKING SYSTEM FOR AIR CARGO SUPPLY CHAIN

Introduction

Previous chapters have analyzed the use of radio frequency identification

technology from the view point of air cargo transportation by exploring a wide variety of

factors such as different materials, frequencies and implementations. The next logical

step would be to describe the advantages radio frequency identification (RFID) brings to

air cargo transportation and utilize these findings to recommend important parameters

of a functional RFID prototype system to be used for air cargo tracking. Even though

there are a few RFID applications in air-cargo business aside from some local solutions

and smaller pilot projects for testing purposes, no major applications yet exist in this

arena (Chang et al., in press).

According to the Merriam-Webster dictionary, a system is "a regularly interacting

or interdependent group of items forming a unified whole" (Merriam Webster). In the

case of general air cargo system, the items are the goods or freight, which can be

grouped into unit load devices (ULDs) whereas the unified whole constitutes the entire

distribution chain. The air cargo distribution chain starts in the cargo warehouse where

the goods are accepted, after which they are loaded into ULDs, brought to the ramp,

loaded onto the aircraft and flown to destination. The chain stops when the goods are

picked-up at the other end. When implementing an RFID tracking system in the air

cargo supply chain, it can be seen as a subsystem of the main transportation system.

Therefore, for the air cargo RFID tracking system, the RFID tags are the items, the ULD

tags represent the group of items and the readers and the infrastructure make the

unified whole.


145









Goal of the system. The global air cargo RFID system should be able to gather

and provide valuable information (items description, weight, dimension and location,

time stamps of freight movement, alarms for temperature abuse, etc) along the

transportation chain in an easy and effortless manner. The placement of RFID tags and

readers should not be intrusive and should help improve operation efficiency and

precision. The idea is to upgrade the quality of information without clogging or

overflowing the databases with useless numbers.

The objective of this chapter is to identify the applicable points of RFID technology

in the air cargo handling process. The main expected benefits would include:

* Ability to track shipments at the item level (shipment visibility)
* Improve operational performance and efficiency (simplify processes, manage
recoveries and decrease processing time)
* Improved customer experience (shipment visibility online)
* Minimize reliance on manual input (reduce claims and performance failure)

However, for the implementation to give the best success it is necessary to evaluate

carefully which technology is best in terms of frequency, type of tags, surrounding

materials, etc. The findings in this dissertation will serve as a guideline when providing

these recommendations. The following sections will describe the potential use case of

RFID for each individual air cargo operation in greater detail.

Typical Air Cargo Warehouse Operations

The general process of freight handling as seen today in an air cargo warehouse

is described in details in tables 7-1 to 7-5, while a brief overview is presented in Figure

7-1. The left columns of the tables describe the current process, whereas the right

column explains how an RFID tracking system could improve the process. The actors


146









taking part are: cargo agent, station attendant, cargo planner, lead agent, booking

coordinator, consignee and shipper.

The process ameliorations described in tables 7-1 to 7-5 can have more positive

consequences than what they were originally designed for. In process #2 (Table 7-1),

tagging every individual piece of a shipment with an RFID tag, instead of tagging every

piece with the exact same label, allows unique identification of each piece as well as

weight management. This is useful further down the road when they are loaded onto

ULDs (process #7, Table 7-2). By knowing the weight of each single item it would be

possible to estimate the ULD total weight more accurately, and therefore give a better

precision to the "weight and balance" team who decides which container goes where in

the cargo hold. In contrast, when a shipment of 10 boxes is accepted today, only the

collective weight of all the boxes is recorded and every box has the same identifying

label (airway bill number, destination, client information, etc.), apart from showing "1 of

10", "2 of 10", "3 of 10", etc. Moreover, knowing the dimension of each single item would

be efficient for the planning agent (process #5, Table 7-2). For example, he could know

ahead of time when an odd shaped package has to be planned onto a pallet because it

would simply not fit inside a container.

ULD/item association is currently done manually and ineffectively by writing down

which shipment is in which ULD (process #7, Table 7-2). Automatic association via

RFID would minimize the risk of human error or hard to read messy hand writing. It

could also be used to tell the station attendant if something is missing or if something

should not be loaded right away. ULD tagging is not only useful for item/ULD

association (process #7, Table 7-2) and ULD movement tracking in and out of the


147









warehouse (processes #11 and #12, Table 7-3), but also allows faster item locating and

easier recovery if they were not placed in the right area. Each year, larger cargo airlines

lose 5-6% of their ULD inventory amounting to hundreds of millions of dollars in loss -

due to breakdowns in their ULD tracking-facilities (Skorna and Richter, 2007).

Typical Air Cargo Ramp Operations

When ULDs and tub carts are brought to the ramp prior to loading of the aircraft,

some unpredictable and undesirable events can occur. For instance, the cargo could be

dropped off at the wrong gate, or conversely, when the aircraft is being unloaded, the

cargo could sit on the tarmac for a long time if the runner got the wrong message or

forgot one of the ULDs at the ramp. Immediate knowledge of "which cargo is where and

when" would contribute to minimization of such mistakes and optimization of the

operations. Moreover, temperature tracking of perishable cargo on the tarmac could

help prioritize the movement of goods by setting off alarms when sensitive products are

being exposed to extreme conditions.

This study showed that even short periods of time on the tarmac could lead to high

temperature elevations at the top of the ULDs (see chapter 6) for summer months. The

opposite is probably true during winter, as the outside temperature is well below

freezing, the freight can face highly damaging environment. Either way, cargo being

exposed to outdoor conditions of all kinds is very vulnerable to temperature abuse and

should be monitored to improve quality control.

Ramp operations are managed by airport employees, so when the air cargo

company delivers the goods to the ramp the managing of the goods is out of their

control until they are unloaded at destination. Goods can be delivered to the ramp up to

2h in advance of flight schedule. If the flight is delayed, the goods can stay on the


148









tarmac for long periods of time. When the weight and balance calculation is ready, the

ground crew can start loading the cargo inside the plane according to the loading plan.

The introduction of an RFID system would allow automatic time stamps and

confirmation that cargo is on board (for operation efficiency and customer information

update). Furthermore, in case a ULD is bumped (not flying on schedule) due to flight

overweight for example, a message could be sent immediately to inform parties of this

situation. Thus, using an RFID tracking system would help reduce the time that

sensitive cargo may spend sitting on the tarmac under diverse weather conditions and

sometime for long periods of time.

In addition, RFID instrumented cargo holds could permit temperature monitoring of

sensitive goods. Or in a simpler application, it could notify the pilot of the recommended

temperature to set the cargo hold at during flight based on the transported goods and

the required temperature range information recorded on the tags. Currently, only

aircraft with temperature control capabilities are equipped with temperature sensors in

their cargo hold (Howard, 2010, personal communication). Temperature monitoring

would not only benefit perishables or live animals; it could also act as a back-up security

system against adverse situations in the cargo hold such as fires.

Findings from this Study and Recommendations for RFID Tracking System
Implementation

RFID implementation necessitates the incorporation of many factors and variables

and the corresponding optimization based on the unique properties of the

implementation environment. Even though the scope of the study discussed in this

dissertation has not included all the aspects required for a successful commercial RFID

implementation, it still provides invaluable information on some of the major variables


149









when implementing an RFID system such as choosing the right tag, the right frequency

and the right ULD material. Following sections will describe these recommendations in

greater detail.

Passive and Active RFID Tags

RFID systems can either be passive, which mean they communicate via signal

backscattering and rely on RF energy transferred from the reader to the tag to power

the tag; or they can be active or semi-active, meaning the tags have their own internal

power source (typically a battery) to continuously power the tag and its RF

communication circuitry, leading to longer read ranges than for passive tags. For the

purpose of simplification, in this text, the word active will be used to represent both

active and semi-active RFID tags.

RFID tags used for piece level identification are only being used once, and

consequently have to be cheap. On the other hand, ULD tags can be permanently

installed and reprogrammable, and therefore justifies a higher cost. From a practical

point of view for air cargo, active tags are more appropriate for ULD identification

because of their bigger size and longer read range. For similar reasons passive tags are

better suited for piece level identification due to their smaller size and affordability.

Moreover, active tags can easily support sensor applications such as temperature

monitoring.

As was observed in chapter 6, cargo encounters wide temperature variations

during transit, hence the need for a temperature monitoring solution. Temperature

control inside the aircraft would be made possible with the use of active RFID

temperature tags. RF propagation characteristics evaluated in this study (see chapter 4)

would permit much better reader-tag communication for active systems than passive


150









systems. To achieve a RFID tag read, two communication links must be successful.

First, the reader-to-tag link must not fail, and second, the tag-to-reader link has to be

complete. In the case of passive systems, the success of the communication is often

limited by the reader-to-tag link (Nikitin and Rao, 2006; Dobkin, 2008). The reader has a

specific sensitivity in the order of -65dBm to -120dBm, whereas the passive tags have

sensitivities of around -12dBm (Nikitin et al., 2009). Therefore, when the reader sends a

signal into its surrounding and the signal is attenuated with distance and interfering

objects if the tags receive enough energy to respond, their signal will most likely be

received with enough energy at the reader end as well. In the case of active RFID

systems, the tags do not need to collect a minimum amount of energy from the reader

to broadcast. Having their own power source allows them to transmit their information

autonomously. As a consequence, it is the "tag-to-reader" link that sets the limitation. In

other words, it is the sensitivity of the reader that will determine the success of the

communication, and therefore permit a much longer read range than for passive

systems.

The signal levels observed in chapter 4 were from -16 to -27dBm for 433MHz, -6

to -17dBm for 915MHz, and -18 to -27dBm for 2.45GHz (Appendix D). If a typical

passive RFID tag has a sensitivity or threshold of around -12dBm (Nikitin et al., 2009),

even 915MHz would not offer enough coverage to read the tags anywhere in the cargo

hold when using this test set-up (one reader antenna). Moreover, this test was

performed inside an empty cargo compartment, which means the signal strengths would

most likely be further reduced in the case of a fully loaded cargo hold. Further testing is

required, but so far active RFID systems are thought to be a more feasible solution.


151









ULD Materials

Air cargo containers or ULDs can be made of different materials. Older ULDs were

all made of aluminum, whereas newer ones are made of Kevlar composite walls on an

aluminum frame. Old containers are being replaced by this new style because of their

much lighter weight (Howard, 2010, personal communication; Nordisk, 2010). This

change is favorable to the findings of this study (see chapter 5), which state that

Kevlar is highly RF-lucent, whereas aluminum is RF-opaque. In other words, Kevlar

lets RF waves go through when aluminum totally reflects them. This result suggest that

most RFID tags applied to the surface of Kevlar UDLs would presumably lead to better

readability than those applied on aluminum ULDs. Moreover, in the case of temperature

monitoring, using Kevlar@ ULD holds a strong advantage over aluminum since their

content information could be read directly through the walls.

Frequency

Warehouse. The warehouse environment is susceptible to many external noises,

such as 2-way radios, wireless networks, cell phones and automatic door entry

systems, which could interfere with the RFID communication links as described in

chapter 3. Based on this study, the frequencies of choice are 433MHz and 915MHz

because of lower interference levels around these bands. Using 915MHz would give the

flexibility of using active and/or passive RFID systems because of the higher power

output allowance by FCC regulations in this band (FCC, 2008); whereas 433MHz is only

suitable for active RFID systems. Using both active and passive systems could be a

plausible solution as well given the use case requirements. As mentioned earlier, item

tags have to be cheap because they are generally not being reused in the system,

whereas container tags can be reprogrammable and permanently installed on the unit.


152









Taking into consideration the interference levels and the flexibility of the systems,

915MHz seems to be the best option for passive RFID, but both 433MHz and 915MHz

could serve for the active. 2.4GHz displays a more significant interference due to

wireless and GSM networks and thus not recommended for warehouse implementation.

Aircraft. The frequency of choice for cargo hold identification was shown to be

915MHz (see Chapter 4) because of an allowed maximum output power higher than at

433MHz (FCC, 2008) as well as a lower attenuation compared with 2.45GHz. As stated

earlier, the higher signal level makes passive RFID systems a possible solution. On the

other hand, when considering active systems, 433MHz might be a good alternative due

to lower attenuation. However, before RFID implementation can take place inside an

aircraft, many studies will have to be conducted to prove that RF signal would not be

significantly interfering with other aircraft radio systems as per the Federal Aviation

Administration document AC20-162 "Airworthiness approval and operational allowance

of RFID systems" (FAA, 2008) which, in the end, could favor the use of passive tags.

International compatibility. ISM bands around 433MHz and 2.45GHz are

available internationally. On the other hand, as much as 915MHz seems to be a good

solution for many different applications, it is only allowed for use in the Americas

("region 2" according to the International Telecommunication Union). The rest of the

world has different regulations for using similar frequencies, which can fall between 860-

960MHz (standard ISO/IEC 18000-6). Therefore, a global air cargo tracking system

would have to account for all those frequencies to be implementable everywhere. There

exists tags that can function anywhere within that range, but the different regulations

can imply different maximum output powers, different bandwidths, etc. Consequently, a


153









system that works in the United States would not necessarily work in an identical

manner elsewhere. Such limitations would need to be taken into consideration in detail

before a global system can be designed with assurance.

Conclusion

This chapter draws upon the results and conclusions of previous chapters to list

the advantages and important parameters of a functioning RFID system from the view

point of air cargo transportation. A detailed comparison of today's air cargo supply chain

and an RFID enabled version is presented to show how greatly RFID would improve the

visibility throughout the entire supply chain. In addition, the findings of previous

chapters such as the interaction of air container materials with RFID, the advantages of

different RFID frequencies in different situations, etc. are utilized to recommend

parameters (such as the type of tag or the frequency band) based on the use case

scenario. It is important to note that a fully functional implementation of RFID in air

cargo supply chain would require full collaboration of many different parties involving

private companies and government institutions. However, when the advantages

presented in this chapter are taken into account, it is trivial to see such efforts would be

beneficial for the entire air transportation industry.


154









Table 7-1. Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo acceptance part).
Cargo acceptance


Improvement with RFID system


# Today's process
1 Cargo agent determines origin and
destination of requested service,
determines compatibility of shipment
with aircraft and station, and if
acceptable, confirms booking with
customer.
2 Station attendant receives freight from
customer and verifies size, weight and
number of pieces at acceptance dock.
Information is hand written on a piece
of paper.







3 Cargo agent verifies documents and
accepts cargo if acceptable. He
creates an airway bill (AWB), prints
and places labels in a tray for station
attendant.
4 Station attendant attaches labels to
shipment and delivers it to build-up
area.


155


Product information is associated with
the RFID tag ID which is applied to every
piece of a shipment. Each piece has its
own weight, dimension and special
requirements (temperature, dangerous
goods, live animal, priority, etc.) info
associated with its RFID tag.

Option: Weighing, dimensioning, tag
programming and application could all be
achieved automatically via a conveyor
belt system.
Tag ID and AWB number are associated
by reading the tag with a handheld RFID
reader.


Labels have already been applied to
shipment.









Table 7-2. Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo build-up part).
Cargo build-up


# Today's process
5 Cargo planner creates build-up plan.
6 Booking coordinator pulls build-up
plan.
7 Station attendant loads shipment as
per build-up plan and creates a
planning load assembly (PLA)
concurrently.


8 Booking coordinator checks for
additional shipments, adds to and
finalizes build-up plan.
9 Station attendant prepares final loads
and PLAs and sends them to the
planners 2h prior to the flight
schedule.
10 Planner enters information in the
database and prints Runner's log.


Improvement with RFID system


Container ULD:
Each ULD has its own RFID tag, which is
scanned simultaneously as the items are
being loaded inside. This creates
automatic association of the item level
pieces and ULDs. This could be
achieved with a wearable or handheld
RFID reader.

Pallet ULD:
Build-up areas for pallet are
predetermined by roller system on the
floor. Therefore, item association could
be done with a fixed RFID reader on the
ceiling above the pallet build-up pit.

Note: All ULD tags include their tare
weight information so that shipment
association gives a precise estimate of
the ULD total weight after build-up.



The information is sent to the planner
automatically from the database as the
ULDs are created.

This step is done automatically since all
cargo information was updated in the
database during build-up.


156









Table 7-3. Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo to/from the ramp section).
Cargo to/from the ramp
# Today's process Improvement with RFID system
11 Station attendant stages and runs All ULD IDs are automatically read when
freight to aircraft 2h prior to flight crossing the warehouse export doors
schedule. (portal RFID reader). Database is
updated of the goods departure.


(Cargo delivered at the ramp is now in the hands of airport employees)


12 Station attendant checks teletype for
any special commodities messages
(for inbound freight), retrieves time
sensitive goods first and delivers to
warehouse and informs Lead agent.




13 Station attendant retrieves and
delivers non time-sensitive goods to
import side of warehouse.

14 Cargo agent (In-flight coordinator)
begins database check-in, prints
inbound manifest and verifies with
physical goods.


Station attendant receives an alarm
when time sensitive goods are ready for
pick-up. An additional message tells him
the required temperature of the item for
proper storage location in warehouse.
All ULDs are automatically read at
warehouse inbound door. Database is
automatically updated of the goods
arrival.
All ULDs were automatically read at
warehouse inbound door. Database is
automatically updated of the goods
arrival.
Already done by automatic reading of the
goods through inbound doorway
entrance.


157









Table 7-4. Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo break-down and storage section).
Cargo break-down and storage


# Today's process
15 Station attendant sends copy of
inbound manifest to lead agent,
retrieves non shipper loaded units
from storage, breaks down and sorts
by AWB, moves to appropriate
location, scans notes location on AWB
and returns completed manifest to
cargo agent.
16 Station attendant moves shipper
loaded units (SLUDs) and connecting
shipments to proper location.
17 Cargo agent checks-in SLUDs in
database.
18 Cargo agent check-in goods not
previously scanned, determines if
pieces or entire shipments are
missing, performs missing cargo
transaction and completes check-in
process.
19 Cargo agent identifies perishable, Live
or hold for pick-up, contacts customer
electronically or by phone, records
conversation, completes paperwork
and waits for customer to contact
back. If no contact within 14 days
warning is mailed to consignee. If no
contact within 30 days, final warning is
mailed to consignee and shipment is
destroyed if destination is domestic or
reported if not domestic.
20 Cargo agent determines if customer
requires delivery. If "no" and goods
are perishable, live or priority, cargo
agent informs customer of 24h pick-up
window, otherwise informs customer
of 48h pick-up window. If "yes",
arranges for ground transportation as
per customer priority.


Improvement with RFID system
During break-down, every piece is
disassociated from the ULD, the same
way it was associated previously.
Pieces are read and associated with their
storage location.



ULD tags should be read and associated
with their waiting location.

Automatic with scanning in previous
step.
Missing cargo should be triggered
automatically since all goods were read
at the inbound door.



Alarm is sent to cargo agent when time
sensitive shipments have entered the
warehouse.
Customer receive an automatic email
when their goods have arrived at the
warehouse and are ready for pick-up.





Automatic from database through email.


158









Table 7-5. Current processes as well as proposed RFID solutions for the air cargo
supply chain (cargo delivery part).
Cargo delivery


Improvement with RFID system


# Today's process
21 Cargo agent delivers documents to
broker or consignee if delivery is
requested and goods are not
domestic, for Customs clearance.
Once cleared, cargo agent checks
AWB for Customs stamp.
22 Cargo agent obtains consignee
signature on AWB and collects
outstanding charges for goods picked
up, completes delivery process in
database and identifies warehouse
location on AWB, stamps "OK to
release" on AWB and delivers AWB to
consignee.
23 Station attendant receives AWB from
consignee, checks for "OK to release"
stamp and retrieves shipment.
24 Station attendant ensures that
consignee inspects shipment for
damage before delivery and releases
to if not damaged. If damaged,
completes bad order report and attach
to AWB.


Cargo eiand met JILDbild e reighl Loa
Acceptance 1 Dimension build? up o ramp
*II x I Faircra fl


FIh Unloading Bring freight Srage
h.,p Fiyhl amval Fre.~hl from -- back ID --.Unload ULD Slorage
eparture ai warehHuse |I /

CUstIOrrir
niolication Cuslomer
idocs pick up

Figure 7-1. Overview of the air cargo operations where the circled steps represent
suggested RFID reading points.


159


Before goods are released, an RFID
read is required to update the database.

Final database update, case closed.









CHAPTER 8
GENERAL CONCLUSION

The goal of this dissertation was to explore the possibility of using radio frequency

identification (RFID) to improve air cargo operations in terms of efficiency, safety and

monitoring. This study showed interference levels at three ultra high frequencies (UHF)

recorded in two air cargo warehouses. The interference levels from highest to lowest

were at 2.45GHz, 433MHz and 915MHz respectively. There are ways to filter out

interference when designing an RFID system, and most of today's high end readers

have that feature. However, the best way to avoid possible disturbing noises is surely to

install the RFID system in an interference free environment. According to the results of

these tests, which were performed inside two warehouses which may or may not give a

realistic representation of all air cargo warehouses in the world, implementation of RFID

systems at 915MHz in North America would bring the best results, interference-wise.

This study also demonstrated that frequencies have a major influence on signal

propagation, especially inside a metal environment. Lower frequencies suffer less

attenuation over distance, but have higher variation within the cargo hold. It was also

demonstrated that antenna polarization can have a significant effect on signal

propagation in some cases, and therefore should not be omitted when designing an

RFID system for air cargo transportation. Moreover, FCC regulations restricts output

powers at 433MHz more than at 915MHz and 2.45GHz, leading to the conclusion that

more RF energy would be available in the cargo hold for reader/tag communication at

915MHz than at the other frequencies tested. Moreover, the study showed that the

relationship between signal strength and tag reads is an important tool to take into

account when implementing RFID systems.


160









This dissertation verified the effects of five commonly used air cargo container wall

materials on RF propagation at the same three different frequencies. Three different

tests were utilized to analyze the characteristics of RF propagation for each material

and the results from all experiments showed a very strong effect of aluminum on RF

transmission and minimal interaction for all other sample materials as expected. These

finding suggest that the use of non-metallic containers for air transportation of

perishable products should make real time temperature monitoring possible by allowing

RF waves to transmit through the wall surface effortlessly.

This study demonstrated that a major temperature gradient can be found within

the same ULD during ground operations as well as during flight, especially when the

flight time exceeds 4h. Therefore, it is suggested that temperature sensitive shipments

should be placed accordingly inside the ULD. This test was performed in the summer,

for that reason, the increase of temperature during ground operations should be

considered variable. However, the temperature distribution observed during flight should

be consistent through the year since temperatures at high altitude do not vary widely.

This work presented a detailed comparison of today's air cargo supply chain and

an RFID enabled version to show how greatly RFID would improve the visibility

throughout the entire supply chain. It is important to note that a fully functional

implementation of RFID in air cargo supply chain would require full collaboration of

many different parties involving private companies and government institutions.

However, when the advantages presented in this dissertation are taken into account, it

is trivial to see such efforts would be beneficial for the entire air transportation industry.


161












APPENDIX A
DC-10 CARGO HOLD AND CARGO DOOR SPECS


66- BY 70-IN. OR


24 HALF-WIDTH CONTAINERS;
158 CU FT EACH (4.47 M3)
3792 CU FT TOTAL (107.38 M3)


GROSS WEIGHT
7000 LB EACH
(3175 KG)


TARE WEIGHT
600 LB EACH
(272.2 KG)


LD3 CONTAINER
GROSS WEIGHT
3500 LB EACH
(1588 KG)


Figure A-1. Standard cargo compartment and containers for model DC-10 series 10,
10CF, 30, 30CF, 40 and 40CF (Boeing, 2010c).


162


TARE WEIGHT
320 LB EACH
(145.1 KG)












STA 949


rnr--_ PLAN


=- APL NOSE




104 IN. (264.2 CM)


A





FRP ELEVATION
15.9 IN.
66 INCHES 44IN
(167.6 CM) 44 IN.
t 1 (111.8 CM)
SEE SECTION 2.3
FOR GROUND CLEARANCE DOR W ACTC COROPANEL






211.3 IN.
CONSTANT SECTION (536.7 CM)
DIAM = 237 IN. --1350 FULL OPEN
(602.0 CM)

FRP 89.8 IN.
18 IN. (228.1 CM)
(45.7 CM)/ I


SECTION AA CRITICAL CLEARANCE LIMIT
19.7 IN.
LOOKING FORWARD C)
(50.0 CM)



Figure A-2. Forward cargo loading door, model DC-10 series 10, 10CF, 30, 30CF, 40
and 40CF (Boeing, 2010c).


163









APPENDIX B
RADIO FREQUENCY ATTENUATION SURFACE PLOTS

All plots are following the color coded spectrum where red is high attenuation, or

70dBm (weak signal) and purple is low attenuation, or 30dBm (strong signal) as

indicated in the example graph below. Slices are numbered from 1 to 12 which

represent the distance from the front of the cargo hold in meters. Surface plots are

shown as if you were standing at the back of the aircraft, looking forward. Therefore, the

right side of the plot is the starboard side of the vessel and the left is port, as also

indicated in the example plot below.






High
.) 35
(D 40

o 50
I 1 55
a- u-- [60
65
70
Center Starboard

Tripod sideways position inside the cargo hold


Figure B-1. Attenuation surface plot example one slice of data (dBm)


164









.! 7 !





2 1 1


3 m


11 8 10




5 : 11 :



6 : 12 1:



Figure B-2. Attenuation surface plot for 433MHz, top end antenna position and circular
polarization.






165








m:
S--"






I









5 II m*
m:-










circular polarization.
circular polarization.


166










110




2 1 .





F'. Iau



i: I:





5 11
nI




6I I12



Figure B-4. Attenuation surface plot for 915MHz top end antenna position and circular
polarization.








167








1 1 7 I *





2 : i

3 9



4 10 ,
m" '
A -M













Figure B-5. Attenuation surface plot for 915MHz, top end antenna position and linear
polarization.





168












!1* 0





3 i: 9









14



*H 2 :--,
I, II,


Figure B-6. Attenuation surface plot for 915MHz, center ceiling antenna position and
circular polarization.







169









a


U:
I
**'


31-


* I
i: 11


5



6.


I:
inB

,fl:


I,
i:


Figure B-7. Attenuation surface plot for 2.45GHz, top end antenna position and circular
polarization.





170


I44


I











3!:
IU



l'

i:'


~~<


I
I'


9


i ': 11
i:
I. j
*


Figure B-8. Attenuation surface plot for 2.45GHz, top end antenna position and linear
polarization.


171


1. 7
I
i I:


r








1


2



3



4



5
o!!


i.
U:
I'

I':

U
I


I1'
I :'


I..


Figure B-9. Attenuation surface plot for 2.45GHz, center ceiling antenna position and
circular polarization.





172


I.
n"it


I









APPENDIX C
STATISTICAL ANALYSIS RESULTS FOR DC-10 RADIO FREQUENCY
PROPAGATION

All statistical analyses were computed using SAS 9.1 (SAS Institute Inc., Cary NC)

and significance was accepted at level a = 0.01.

Table C-1. Effects of frequency, antenna location and antenna polarization on
attenuation levels of the complete dataset.
Effects mean F value p-value
Frequency
433MHz 41.69
915MHz 47.64 1258.26 < 0.0001
2.45GHz 59.34
Location
Top End 51.39
T E 5139 15.19 < 0.0001
Ceiling 49.12
Polarization
Circular 49.23
C lar 4923 68.81 < 0.0001
Linear 54.47


173









Table C-2. Effect of width on
and polarization.


attenuation levels for each frequency, antenna location


Constants
Frequency Location Polarization


433


Top End


Top End


Ceiling


Circular


Linear


Circular


Effects
Width
Port
Center
Starboard
Port
Center
Starboard
Port
Center
Starboard


mean F value p-value


42.63
42.03
40.45
N/A
N/A
N/A
42.51
41.22
41.30


2.06 0.1328


N/A


N/A


0.84 0.4343


Port 48.36
Top End Circular Center 47.57 0.46 0.6331
Starboard 48.06
Port 48.85
Top End Linear Center 47.38 0.62 0.5417
Starboard 47.84
Port 46.94
Ceiling Circular Center 46.55 0.39 0.6760
Starboard 47.26


Port
Center
Starboard
Port
Center
Starboard
Port
Center
Starboard


58.77
58.32
57.87
62.60
58.39
61.75
58.11
58.47
59.77


0.4992


12.5 < 0.0001


3.04 0.0522


174


915


2450


Top End


Top End


Ceiling


Circular


Linear


Circular









Table C-3. Effect of height on attenuation levels for each frequency, antenna location
and polarization.
Constants Effects mean F value p-value
Frequency Location Polarization Height


High
Middle
Low
High
Middle
Low
High
Middle
Low


41.51
41.43
42.16
N/A
N/A
N/A
41.11
41.74
42.18


0.25 0.7770


N/A


N/A


0.46 0.6353


High 48.43
Top End Circular Middle 47.01 2.16 0.1210
Low 48.54
High 52.66
Top End Linear Middle 46.37 27.13 < 0.0001
Low 45.03
High 46.86
Ceiling Circular Middle 46.94 0.01 0.9936
Low 46.94


High
Middle
Low
High
Middle
Low
High
Middle
Low


58.25
58.87
57.85
61.91
60.28
60.54
59.16
58.92
58.28


0.94 0.3943


1.61 0.2054


0.4524


175


Top End


Top End


Ceiling


Circular


Linear


Circular


433


915


2450


Top End


Top End


Ceiling


Circular


Linear


Circular









Table C-4. Effect of depth on
and polarization.


attenuation levels for each frequency, antenna location


Constants
Frequency Location Polarization


Top End


433MHz


Top End


Ceiling


Circular













Linear













Circular


Effects
Depth
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12


F value p-value


mean
(dBm)
36.61
38.97
37.73
39.65
39.15
42.32
43.51
43.30
44.41
43.47
44.21
47.11
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
46.96
44.10
40.47
39.97
40.82
36.92
39.03
40.08
42.96
43.24
42.58
42.93


N/A


N/A


3.61 0.0003


176


6.28 < 0.0001









Table C-4. Continued
Constants Effect Mean F value p-value
Frequency Location Polarization Depth (dBm)


1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12


43.63
45.48
45.99
46.45
47.21
47.67
47.94
48.72
48.28
50.95
52.67
50.94
42.22
44.54
44.94
45.02
45.40
47.32
49.07
49.64
52.04
51.61
52.48
51.94
50.81
49.51
46.75
47.27
44.68
43.25
42.37
45.15
46.86
47.35
43.59
50.32


Top End


915MHz


Top End


Circular













Linear













Circular


8.33 < 0.0001













4.88 < 0.0001













11.82 < 0.0001


Ceiling


177









Table C-4. Continued
Constants
Frequency Location


Top End


2.45GHz


Top End


Ceiling


Polarization


Circular













Linear













Circular


F value p-value


Effect
Depth
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12


178


Mean
(dBm)
54.57
55.44
56.77
58.37
57.55
57.78
58.29
59.38
59.09
59.96
60.69
61.90
60.05
60.70
60.81
59.82
60.65
61.07
59.12
61.41
60.61
61.57
62.92
62.17
59.99
59.55
58.61
57.56
56.31
54.82
55.94
58.40
59.19
61.16
61.77
62.08


5.83 < 0.0001













0.53 0.8819













9.95 < 0.0001









APPENDIX D
RADIO FREQUENCY SIGNAL STRENGTH PROPAGATION SURFACE PLOTS

All plots are following the color coded spectrum where purple is high signal

strength, or OdBm and red is low signal strength, or -40dBm as indicated in the example

graph below. Slices are numbered from 1 to 12 which represent the distance from the

front of the cargo hold in meters. Surface plots are shown as if you were standing at the

back of the aircraft, looking forward. Therefore, the right side of the plot is the starboard

side of the vessel and the left is port, as also indicated in the example plot below.


High


Middle


Low


Port


Center


i i +


Starboard


-5
-10
-15
-20
L -25
- -30
--35
-40


Tripod sideways position inside the cargo hold


Figure D-1. Signal strength surface plot example one slice of data (dBm)


179












I
-i


11


1!.


Figure D-2. Signal strength surface plot for 433MHz, top end antenna position and
circular polarization.





180


12










i o:,
go I-











1- 7







5 11 1




6 12



Figure D-3. Signal strength surface plot for 433MHz, center ceiling antenna position
and circular polarization.






181











1 12




I I
2 !



















18210
1. I.

11 --




n.. 12 I.



Figure D-4. Signal strength surface plot for 915MHz, top end antenna position and
circular polarization.






182











1 12






2 18





U. .





5 11 --




6 12



Figure D-5. Signal strength surface plot for 915MHz, top end antenna position and
linear polarization.






183










1 1




2 18 4








U. .
10-.













Figure D-6. Signal strength surface plot for 915MHz, center ceiling antenna position
and circular polarization.






184









I
-i


11


1!.


Figure D-7. Signal strength surface plot for 2.45GHz, top end antenna position and
circular polarization.





185


12









r 8
I:


! i
U:
I

U
!.
II1


Figure D-8. Signal strength surface plot for 2.45GHz,
linear polarization.


top end antenna position and


186


2 1



3


8


10


5


I i
II


11


12


1 I









'I
rl;1


12 o)


9


10



111


Figure D-9. Signal strength surface plot for 2.45GHz, center ceiling antenna position
and circular polarization.





187


-I





-I



I


3:
I




I


3:
Ji,










APPENDIX E
UNIT LOAD DEVICES, TEMPERATURE GRAPHS AND POSITIONS

AKH 2084


50
45
40

35
o, 30
30
A 25
A
S20
E
15
10
5


YYZ-YUL


YUL-YYZ


I-

- -[


I I

I I




BP
I I
I I


Time (h)
Time (h)


AKH 9778


YYZ-YUL


YUL-YYZ


c- --b c- O- -t-


Time (h)

Figure E-1. Temperature profiles of all tags for ULDs (A) AKH 2084 and (B) AKH 9778
to and from Montreal (YUL). First flight segments are highlighted in yellow
and returning flight segments are highlighted in blue. Corresponding container
positions are shown on the A320 sketch to the right.


188


50
45
40
S35
0
o, 30
3
B 25
| 20
E
-15
10
5
0


LI








AKH 1987


YYZ YVR


YVR YYZ


li


L I
Eli

HF
l
O I
O


- --- Time- (J J C h)C
Time (h)


Figure E-2. Temperature profiles of all tags for ULDs AKH 1987 to and from Vancouver
(YVR). First flight segment is highlighted in yellow and returning flight
segment is highlighted in blue. Container position was the same for both flight
and is consequently shown in green on the A321 sketch to the right.


189









AKE 03782


YYZ- LHR


LHR YVR


Time (h)


AKE 04090


YYZ LHR LHR YYZ


B
0.
E
I_


EOL

OE
EO

-7



En
E-n

E]
OOI
OW


0 Co iC' O- O o o t- L0oc L
0 0 0 bbb~~~


Time (h)

Figure E-3. Temperature profiles of all tags for ULDs (A) AKE 03782, (B) AKE 04090,
(C) AKE 04969, and (D) AKE 05335 to and from London (LHR). First flight
segments are highlighted in yellow and returning flight segments are
highlighted in blue. Corresponding container positions are shown on the A330
sketch to the right.


190


A 0
A
E
I_


00-
LIJ
.-1
0
OJ











AKE 04969


YYZ- LHR


LHR-YUL


Time (h)


AKE 05335


YYZ- LHR


-A
-B
-c
-D
E
-F
-G
-H


LHR YUL


_aI D


20
15 -
10
5
0


Time (h)


Figure E-3. Continued.


191


o


C

E


1H





D:3


W1





v1


I
OE

WE
W0







Wl
El--]


50
45
40

-35
2 30
3
25 O


D
E.
E









AKE 03748


YYZ- FRA


o

A
0.
E
I_


FRA- YUL


U
5

0o 11- 1- M W.'.t 0 CD- W M M,,. -
'- r- CD CIN C9 CL CD M T
Time (h)


AKE 04632


YYZ- FRA FRA- YUL


o

B
E
0^


Time (h)

Figure E-4. Temperature profiles of all tags for ULDs (A) AKE 03748, (B) AKE 04632,
(C) AKE 05168, and (D) AKE 05255 to and from Frankfurt (FRA). First flight
segments are highlighted in yellow and returning flight segments are
highlighted in blue. Corresponding container positions are shown on the B777
sketch to the right.


192


Hn







ii
001
hF
88
BBl


On





O,



LD.

HF
001


11c c ) in 1- M^ Wo 'It oCD WM Mo Mo ) in 1- M^ Wo 'It oCD WM Mo Mo U)
*'-~~ ~ IN IN IN IN IN C M MC MCM MO MO MO MO MO COC qt qq -t qqt qqt U) U
0 0 ~~0~0~~0 ~
.............................
~~~~~0 0 0060 0


'- r ,,CO CO 't o (0 CMC CO M0 1- 0
ib0 D 00 '- 0) M -t bD r,.- 0) 0 I


;-;- ---;---









AKE 05168


YYZ- FRA


FRA- YUL


Time (h)


AKE 05255


YYZ- FRA


0


0
D
E
I-


V


c0cir-c


-A
-B
-C
- [I
-F
G
-H


FRA- YUL


r-c~r--c~r--


Time (h)


Figure E-4. Continued.


193


C|
E
I_


00












HI

00
OO


DD
OO




O=
BB
BB

OO


I

0





OU
O



ON









LIST OF REFERENCES


Abad, E., F. Palacio, M. Nuin, A. Zarate, D.E. Gonzales, A. Juarros, J. M. G6mez and
S. Marco. 2009. RFID smart tag for traceability and cold chain monitoring of foods:
Demonstration in an intercontinental fresh fish logistic chain. Journal of Food
Engineering 93(4): 394-399.

Air Canada. 2007. Publication 167 B777 Load planning. 5.

Air Canada. 2006. Publication 163 A330 Load planning. 5.

Air Canada. 2005. Publication 169 A320 Load planning. 6.

Air Canada. 2005. Publication 172 A321 Load planning. 5.

Airbus. 2009. Airbus continues to play major role in composites revolution. Airbus 40Y
of Innovation. Available at:
http://www.airbus.com/en/corporate/innovation/40years-innovation/40-years-
innovation-
items/?tx ttnews%5Btt news%5D=233&tx ttnews%5BbackPid%5D=782&cHash=
e0bce8b6c7. Accessed 05/23 2010.

Altunbas, A. E. 2010. Wearable RFID reader and compact fractal patch antenna design.
Gainesville, Fla.: University of Florida.

Anonym. 2007. Composite materials in aviation. The Airplanes Channel. Available at:
http://www.airplanes.com/blog/composite-materials-in-aviation/. Accessed 04/09
2010.

Anritsu. 2009. Fundamentals of interference in wireless networks. 11410-005361-16.

Anritsu. 2008. Guide to Spectrum Analyzers.

Ashford, N., C. A. Moore and H. P. M. Stanton. 1984. Airport operations. New York:
Wiley.

Bilstein, R. E. 1994. Flight in America: from the Wrights to the astronauts. Rev. ed.
Baltimore: Johns Hopkins University Press.

Blaunstein, N. and C. G. Christodoulou. 2007. Radio propagation and adaptive
antennas for wireless communication links: terrestrial, atmospheric and
ionospheric. Hoboken, N.J.: Wiley-lnterscience.

Boeing. 2010a. Boeing 787 dreamliner will provide new solutions for airlines,
passengers. Boeing. Available at:
http://www.boeing.com/commercial/787family/background.html. Accessed 04/09
2010.


194









Boeing. 2010b. Commercial airplanes. Boeing. Available at:
http://www.boeing.com/commercial/safety/index.html. Accessed 04/09 2010.

Boeing. 2010c. Commercial airplanes DC-10 family. Boeing. Available at:
http://www.boeing.com/commercial/dc-10/index.html. Accessed 04/25 2010.

Boeing. 2008. World air cargo forecast 2008-2009. Accessed 09/02 2009.

Bollen, A. F., D. W. Brash and B. L. Bycroft. 1998. Air-freight coolchain improvements
using insulation and supplemental cooling. Applied Engineering in Agriculture
14(1): 49-53.

Catto-Smith, C. 2006. Thermal protection of perishable air cargo during ground
handling. In International symposium on fresh produce supply chain management.

Cerino, A. and W. P. Walsh. 2000. Research and application of radio frequency
identification (RFID) technology to enhance aviation security. In Proceedings of
the IEEE 2000 National Aerospace and Electronics Conference. NAECON 2000,
127-135.

Cerino, A., G. L. Fuller and R. W. Devereux. 1997. EMI from sources in the cargo bay.
In Digital Avionics Systems Conference, 1997. 16th DASC., AIAA/IEEE, 4.1-9-16
vol.1.

Chang, Y. S., C. H. Oh, Y. S. Whang, J. J. Lee, J. A. Kwon, M. S. Kang, J. S. Park and
Y. P. Ung. 2006. Development of RFID enabled aircraft maintenance system. In
2006 IEEE International Conference on Industrial Informatics, 224-229.

Chang, Y. S., M. G. Son and C. H. Oh. Design and implementation of RFID based air-
cargo monitoring system. Advanced Engineering Informatics. In Press, Corrected
Proof.

Chen, S. C. Q. and V. Thomas. 2001. Optimization of inductive RFID technology. In
Proceedings of the 2001 IEEE International Symposium on Electronics and the
Environment, 82-87.

Cheng, D. K. 1993. Fundamentals of engineering electromagnetics. Prentice Hall.

Chow, H. K. H., K. L. Choy, W. B. Lee and K. C. Lau. 2006. Design of a RFID case-
based resource management system for warehouse operations. Expert Systems
with Applications 30(4): 561-576.

Clampitt, H. G. 2006. RFID certification textbook. 2nd ed. Houston, TX: PWD Group,
Inc.


195









Day, D. A. 2009. Composites and advanced materials. U.S. Centennial of flight
commission. Available at:
http://www.centennialofflight.gov/essay/EvolutionofTechnology/composites/Tech
40.htm. Accessed 01/26 2010.

Devereux, R. W., G. L. Fuller and R. Schillinger. 1997. Further analysis of the CV-580
and B-727 aircraft RF coupling experiment data. In Digital Avionics Systems
Conference, 1997. 16th DASC., AIAA/IEEE, 4.1-1-8 vol.1.

Devereux, R. W., G. L. Fuller and R. Schillinger. 1997. Electromagnetic susceptibility of
installed avionics. In Digital Avionics Systems Conference, 1997. 16th DASC.,
AIAA/IEEE, 4.1-17-24 vol.1.

Dobkin, D. M. and S. M. Weigand. 2005. Environmental effects on RFID tag antennas.
Microwave Symposium Digest, 2005 IEEE MTT-S International.

Dobkin, D. M. 2008. The RF in RFID: passive UHF RFID in practice. Amsterdam;
Boston: Elsevier / Newnes.

Domdouzis, K., B. Kumar and C. Anumba. 2007. Radio-Frequency Identification (RFID)
applications: A brief introduction. Advanced Engineering Informatics 21(4): 350-
355.

Elias, B. 2007. CRS report for congress Air cargo security. RL32022.

Ely, J. J., T. X. Nguyen, S. V. koppen, M. T. Salud and J. H. Beggs. 2003. Wireless
phone threat assessment and new wireless technology concerns for aircraft
navigation radios. NASA/TP-2003-212446.

Emond, J. P. 2008. The cold chain. In RFID Technology and Applications, 144. ed. S. B.
MilesSarma, S. E. and Williams, J. R., New York, NY, USA: Cambridge University
press.

Emond, J. P. 2007. Quantifying RFID's cold chain benefits. In RFID Journal LIVE! 2007
Conference.

Emond, J. P., F. Mercier and M. C. N. Nunes. 1999. In-flight temperature conditions in
the holds of a widebody aircraft. In Proceedings of the 20th InternationalCongress
of Refrigeration, Paper No. 281.

ETSI. 2007. Electromagnetic compatibility and Radio spectrum Matters (ERM);
Technical characteristics of RFID in the UHF Band; System Reference Document
for Radio Frequency Identification (RFID) equipment; Part 1: RFID equipment
operating in the range from 865 MHz to 868 MHz. ETSI TR 102 649-1.

FAA. 2008. Airworthiness approval and operational allowance of RFID systems. AC20-
1621.


196









FAA. 2006. FAA aerospace forecast fiscal years 2006-2017. 1-90.


FCC. 2008. Part 15 Radio frequency devices. 1-157.

Finkenzeller, K. 2003. RFID handbook: fundamentals and applications in contactless
smart cards and identification. New York, NY, USA: John Wiley & Sons, Inc.

Flightglobal. 2009. Aircraft profile: Airbus A320. Available at:
http://www.flightglobal.com/articles/2008/01/01/221129/airbus-a320-aircraft-
profile.html. Accessed 09/02 2009.

Forster, P. W. and A. C. Regan. 2001. Electronic integration in the air cargo industry: an
information processing model of on-time performance. Transportation Journal
40(4): 46-61.

Foster, P. R. and R. A. Burberry. 1999. Antenna problems in RFID systems. lEE
Colloquium on RFID Technology 3/1-3/5.

Friis, H. T. 1946. A note on a simple transmission formula. Proceedings of the I.R.E.
34(5): 254.

Fuschini, F., C. Piersanti, F. Paolazzi and G. Falciasecca. 2008. Analytical approach to
the backscattering from UHF RFID transponder. Antennas and Wireless
Propagation Letters, IEEE 733-35.

Griffin, J. D., G. D. Durgin, A. Haldi and B. Kippelen. 2006. RF tag antenna performance
on various materials using radio link budgets. Antennas and Wireless Propagation
Letters, IEEE 5(1): 247-250.

Hallowell, S. F. and P. Z. Jankowski. 2005. Transportation security technologies
research and development. In IEEE Military Communications Conference, 2005.
MILCOM 2005. 1753-1756 Vol. 3.

Henderson, T. 2010. Static electricity Lesson 3. The Physics Classroom. Available at:
http://www.physicsclassroom.com/class/estatics/u813c.cfm. Accessed 07/03 2010.

IATA. 2002. ULD technical manual, 17th edition. Montreal, Canada: International Air
Transport Association.

Impinj. 2007. UHF Gen 2 RFID readers Five factors for success.

Intelleflex. 2010. The Intelleflex XC3 technology platform: An implementation of the new
ISO Class 3 standard.

Jacinto, C., M. Romano, A. Montes, P. Sousa and M. S. Nunes. 2009. RFID tuning
methodology applied on airport baggage tracking. In IEEE Conference on
Emerging Technologies & Factory Automation, 2009. ETFA 2009. 1-4.


197









Jedermann, R. and W. Lang. 2007. Semi-passive RFID and beyond: steps towards
automated quality tracing in the food chain. Int. J. Radio Frequency Identification
1(3): 247-259.

Jedermann, R., J. P. Emond and W. Lang. 2007. Shelf life prediction by intelligent RFID
technical limits of model accuracy. In International Conference on Dynamics in
Logistics.

Jedermann, R., L. Ruiz-Garcia and W. Lang. 2009. Spatial temperature profiling by
semi-passive RFID loggers for perishable food transportation. Computers and
Electronics in Agriculture 65(2): 145-154.

Keskilammi, M., L. Sydanheimo and M. Kivikoski. 2003. Radio frequency technology for
automated manufacturing and logistics control. Part 1: Passive RFID systems and
the effects of antenna parameters on operational distance. The International
Journal of Advanced Manufacturing Technology 21(10): 769-774.

Ketzenberg, M. E. and J. M. Bloemhof-Ruwaard. 2009. The value of RFID technology
enabled information to manage perishables. ERIM Report Series Reference No.
ERS-2009-020-LIS.

Khan, M. A., M. Sharma and P. R. Brahmanandha. 2009. A survey of RFID tags.
International Journal of Recent Trends in Engineering 1 (4).

Kim, D. Y., H. G. Yoon, B. J. Jang and J. G. Yook. 2008. Interference analysis of UHF
RFID systems. Progress in Electromagnetics Research B 4115-126.

Kumar, R., S. K. Kaura, A. K. Sharma, D. P. Chhachhia and A. K. Aggarwal. 2007.
Knife-edge diffraction pattern as an interference phenomenon: An experimental
reality. Optics & Laser Technology 39(2): 256-261.

LaBerge, E. F. C. 2007. An analysis of the effects of RFID tags on narrowband
navigation and communication receivers.

Lahiri, S. 2006. RFID sourcebook. Upper Saddle River, NJ, USA: IBM Press.

Landt, J. 2005. The history of RFID. Potentials, IEEE 24(4): 8-11.

Laniel, M., J. P. Emond and A. E. Altunbas. 2009. RFID behavior study in enclosed
marine container for real time temperature tracking. Sensing and Instrumentation
for Food Quality and Safety 3(1): 34-40.

Liang, B., A. Hu and Z. Qin. 2006. Trends and brief comments on anti-collision
techniques in radio frequency identification system. In 2006 6th International
Conference on ITS Telecommunications Proceedings, 241-245.

Littell, R. C., G. A. Milliken, W. W. Stroup, R. D. Wolfinger and 0. Schabenberber. 2006.
SAS for mixed models. 2nd ed. Cary, NC, USA: SAS Publishing.


198









Longhurst, R. S. 1967. Geometrical and physical optics. 2nd ed. London: Longman.

Mac Carthy, U. M. 2009. Assesment of UHF RFID as an automatic ID data carrier for
item level food supply chain traceability. PhD diss. Dublin: University College
Dublin.

Mayer, L. W., M. Wrulich and S. Caban. 2006. Measurements and channel modeling for
short range indoor UHF applications. In First European Conference on Antennas
and Propagation. EuCAP 2006. 1-5.

McCarthy, T. M. 2003. Conducting a market opportunity analysis for air cargo
operations. Transportation Journal: May 18, 2010.

McCoy, T., R. J. Bullock and P. V. Brennan. 2005. RFID for airport security and
efficiency. In The lEE Seminar on Signal Processing Solutions for Homeland
Security, 2005. (Ref. No. 2005/11108), 9 pp.

Merriam Webster, I. Merriam-Webster's open dictionary [electronic resource].
Springfield, MA: Merriam-Webster.

Meyers, M., M. Brown, S. Patadia and S. Dua. 2007. Mike Meyers' comptia RFID+
certification passport. New York: McGraw-Hill.

Mo, L. and H. J. Zhang. 2007. RFID antenna near the surface of metal. In 2007
International Symposium on Microwave, Antenna, Propagation and EMC
Technologies for Wireless Communications, 803-806.

Moir, I. and A. G. Seabridge. 2001. Aircraft systems: mechanical, electrical, and
avionics subsystems integration. Reston, VA; Bury St. Edmunds, U.K.: American
Institute of Aeronautics and Astronautics; Professional Engineering Pub.

NFPG. 2008. United Kingdom frequency allocation table.

Nguyen, T. X. and J. J. Mielnik. 2008. Radio frequency compatibility of an RFID tag on
glideslope navigation receivers. In IEEE International Symposium on
Electromagnetic Compatibility. EMC 2008, 1-6.

Nguyen, T. X., J. J. Ely, S. V. Koppen and M. S. Fersch. 2008. RFID transponders' RF
emissions in aircraft communication and navigation radio bands. In 2008
International Symposium on Electromagnetic Compatibility EMC Europe, 1-6.

Nguyen, T. X., S. V. Koppen, L. J. Smit, R. A. Williams and M. T. Salud. 2005. Wireless
phone threat assessment for aircraft communication and navigation radios. In
2005 International Symposium on Electromagnetic Compatibility. EMC 2005, 135-
140 Vol. 1.


199









Nguyen, T. X., S. V. Koppen, J. J. Ely, G. N. Szatkowski, J. J. Mielnik and M. T. P.
Salud. 2007. Small aircraft RF interference path loss. In IEEE International
Symposium on Electromagnetic Compatibility. EMC 2007, 1-6.

Nguyen, T. X., J. J. Ely, K. L. Dudley, S. A. Scearce, M. O. Hatfield and R. E. Ri. 2006.
Passenger transmitters as a possible cause of aircraft fuel ignition in support of
an aircraft accident investigation. In IEEE International Symposium on
Electromagnetic Compatibility. EMC 2006, 228-233.

Nguyen, T. X., S. V. Koppen, J. J. Ely, R. A. Williams, L. J. Smith and M. T. P. Salud.
2004. Portable wireless device threat assessment for aircraft navigation radios.

Nikitin, P. V. and K. V. S. Rao. 2009. LabVIEW-based UHF RFID tag test and
measurement system. IEEE Transactions on Industrial Electronics, 56(7): 2374-
2381.

Nikitin, P. V. and K. V. S. Rao. 2008. Antennas and propagation in UHF RFID systems.
2008 IEEE International Conference on RFID, 277-288.

Nikitin, P. V. and K. V. S. Rao. 2006. Performance limitations of passive UHF RFID
systems. In 2006 IEEE Antennas and Propagation Society International
Symposium, 1011-1014.

Nikitin, P. V., K. V. S. Rao, R. Martinez and S. F. Lam. 2009. Sensitivity and impedance
measurements of UHF RFID chips. IEEE Transactions on Microwave Theory and
Techniques, 57(5): 1297-1302.

Nikitin, P. V., D. D. Arumugam, M. J. Chabalko, B. E. Henty and D. D. Stancil. 2010.
Long range passive UHF RFID system using HVAC ducts. Proceedings of the
IEEE PP(99): 1-7.

Nordisk. 2010. Air cargo containers. Teleflex Incorporated. Available at:
http://www.nordisk-aviation.com/default.aspx?Cat=2&ld=11. Accessed 02/02
2010.

NTIA. 2010. Federal spectrum use summary, 30 MHz 3000 GHz.

NTIA. 2003. United States frequency allocations the radio spectrum.

NTSB. 2010. NTSB Aviation Accident database & synopses. National Transportation
Safety Board. Available at: http://www.ntsb.gov/ntsb/query.asp. Accessed 05/21
2010.

Nunes, M. C. N., J. P. Emond and J. K. Brecht. 2003. Quality of strawberries as affected
by temperature abuse during ground, in-flight and retail handling operations. In
239-246. Acta Hort. (ISHS).


200









Nunes, M. C. N., J. P. Emond, K. V. Chau, M. Rauth, S. Dea and W. Pelletier. 2006.
Effects of in-store conditions on the quality of freshfruits and vegetables. 262.

O'Connor, M. C. 2005. IATA approves UHF for bag tags.

Ott, R. L. and M. T. Longnecker. 2004. A first course in statistical methods. Belmont,
CA, USA: Brooks/Cole Thomson Learning.

Penttila, K., M. Keskilammi, L. Sydanheimo and M. Kivikoski. 2006. Radio frequency
technology for automated manufacturing and logistics control. Part 2: RFID
antenna utilisation in industrial applications. The International Journal of Advanced
Manufacturing Technology 31(1): 116-124.

Poon, T. C., K. L. Choy, H. K. H. Chow, H. C. W. Lau, F. T. S. Chan and K. C. Ho.
2009. A RFID case-based logistics resource management system for managing
order-picking operations in warehouses. Expert Systems with Applications 36(4):
8277-8301.

Prothro, J. T., G. D. Durgin and J. D. Griffin. 2006. The effects of a metal ground plane
on RFID tag antennas. In IEEE Antennas and Propagation Society International
Symposium 2006, 3241-3244.

Rao, K. V. S. 1999. An overview of backscattered radio frequency identification system
(RFID). In 1999 Asia Pacific Microwave Conference, 746-749 vol.3.

Rappaport, T. S. 2002. Wireless communications : principles and practice. 2nd ed.
Upper Saddle River, N.J.: Prentice Hall PTR.

Rappaport, T. S. and C. D. McGillem. 1989. UHF fading in factories. IEEE Journal on
Selected Areas in Communications, 7(1): 40-48.

Raumonen, P., L. Sydanheimo, L. Ukkonen, M. Keskilammi and M. Kivikoski. 2003.
Folded dipole antenna near metal plate. IEEE Antennas and Propagation Society
International Symposium, 1848-851 vol.1.

Reed, R. 2009. Refraction of light. Interactagram.com. Available at:
http://interactagram.com/physics/optics/refraction/. Accessed 06/30 2010.

Reitz, J. R., R. W. Christy and F. J. Milford. 1993. Monochromatic waves in bounded
regions. In Foundations of electromagnetic theory, 441-469. ed. Anonymous,
Reading, Mass.: Addison-Wesley.

Roussos, G. and V. Kostakos. 2009. RFID in pervasive computing: State-of-the-art and
outlook. Pervasive and Mobile Computing 5(1): 110-131.

Ruiz-Garcia, L., P. Barreiro and J. I. Robla. 2008. Performance of ZigBee-Based
wireless sensor nodes for real-time monitoring of fruit logistics. Journal of Food
Engineering 87(3): 405-415.


201









SAS Institute. SAS/STAT9.1 user's guide. Cary, N.C.: SAS Institute, Norwood, Mass.

Saunders, B. 2003. Award-winning blast resistant cargo container at Dubai air show
spotlight.

Schmoetzer, K. 2005. Enhanced cargo monitoring container communication interface.
In International Aircraft Systems Fire Protection Working Group, Airbus, dept.
BCECS4.

Schmoetzer, K. 2001. Aircraft fire detection: Requirements, qualification and certification
aspects. In 12th international conference on automatic fire detection, National
Institute of Standards and Technology.

Shahidi, R. 1995. Link budget analysis: student manual. Arlington, VA: LLC.

Skorna, A. C. H. and A. Richter. 2007. RFID and air cargo. RFID Jounal. Available at:
www.rfidjournal.com/article/articleview/3482/. 2008.

Sydanheimo, L., L. Ukkonen and M. Kivikoski. 2006. Effects of size and shape of
metallic objects on performance of passive radio frequency identification. The
International Journal of Advanced Manufacturing Technology 30(9): 897-905.

Syversen, E., F. J. Pineda and J. Watson. 2008. Temperature variations recorded
during interinstitutional air shipments of laboratory mice. Journal of the American
Association for Laboratory Animal Science 47(1): 31-36.

Taylor, M. J. H. 1989. The aerospace chronology. London: Tri-Service Press.

The Aluminum Association. 2008. Aluminum in aircraft. The Aluminum Association.
Available at:
http://www.aluminum.org/Content/NavigationMenu/Thelndustry/TransportationMar
ket/Aircraft/default.htm. Accessed 01/26 2010.

Valenzuela, R. A., O. Landron and D. L. Jacobs. 1997. Estimating local mean signal
strength of indoor multipath propagation. IEEE Transactions on Vehicular
Technology, 46(1): 203-212.

van de Voort, M. and A. Ligtvoet. 2006. Towards an RFID policy for Europe. DRR-4046-
EC.

Vega, H. 2008. Air cargo, trade and transportation costs of perishables and exotics from
South America. Journal of Air Transport Management 14(6, Sp. Iss. SI): 324-328.

Veronneau, S. and J. Roy. 2009. RFID benefits, costs, and possibilities: The
economical analysis of RFID deployment in a cruise corporation global service
supply chain. International Journal of Production Economics 122(2): 692-702.


202









Villeneuve, S., M. O. Nagdi and J. P. Emond. 2001. Heat transfer in air cargo unit load
devices. International Society for Horticultural Science Acta Hortuculturae 566245-
250.

Villeneuve, S., F. Mercier, W. Pelletier, M. O. Nagdi and J. P. Emond. 2000. Effect of
environmental conditions in air shipment of perishables during ground operations.
In ASAE Annual International Meeting, ASAE.

Villeneuve, S. 2006.
Etude des phenomenes d'echange thermique et des conditions environmentales
ayant un impact sur la conservation des denrees perissables lors des operations
aeroportuaires. PhD diss. Quebec: Universite Laval.

Visich, J. K., S. Li, B. M. Khumawala and P. M. Reyes. 2009. Empirical evidence of
RFID impacts on supply chain performance. International Journal of Operations &
Production Management 29(11-12): 1290-1315.

Wang, H., S. Chen and Y. Xie. 2010. An RFID-based digital warehouse management
system in the tobacco industry: a case study. Internation Journal of Production
Research 48(9): 2513-2548.

Wegener, P. P. 1997. What makes airplanes fly?: history, science, and applications of
aerodynamics. 2nd ed. New York: Springer.

Wilmes, H., B. Kolesnikov, A. Fink and C. Kindervater. 2002. New design concepts for a
CFRP fuselage.

Winder, S. and J. J. Carr. 2002. Radio and RF engineering pocket book. 3rd ed. Oxford:
Newnes.

Wright, T. 2008. Temperature sensitive pharmaceutical transportation a changing
picture.

Wu, N. C., M. A. Nystrom, T. R. Lin and H. C. Yu. 2006. Challenges to global RFID
adoption. Technovation 26(12): 1317-1323.

Yonemoto, N., K. Yamamoto, K. Yamada and T. Hirata. 2007. RF emission
measurement of 433 MHZ RFID tags for EMI evaluation to onboard instruments of
aircraft. In 7th International Symposium on Electromagnetic Compatibility and
Electromagnetic Ecology, 232-235.

Yuanxin, O., H. Yao, P. Ling, W. Dongwei and X. Zhang. 2008. An intelligent RFID
reader and its application in airport baggage handling system. In 4th International
Conference on Wireless Communications, Networking and Mobile Computing,
2008. WiCOM '08. 1-4.


203









Zhang Q., G. Cheng, Z. Wang, J. Wei and D. Yan. 2009. Development of RFID
application system in cargo inbound and outbound. In TENCON 2009 2009 IEEE
Region 10 Conference, 1-6.

Zweig, S. E. 2006. Advances in vaccine stability monitoring technology. Vaccine 24(33-
34): 5977-5985.


204









BIOGRAPHICAL SKETCH

Magalie Laniel was born in Montreal, Quebec, Canada. She attended Laval

University in Quebec city where she received, in 2004, a Bachelor of Engineering in

food engineering. In the following fall, she began a master's under the direction of Dr

Emond in the Department of Agricultural and Biological Engineering at the University of

Florida. Her masters' work opened the path to continue her studies towards the PhD.

Along the years, she has been working on several projects involving radio frequency

identification (RFID) technology, packaging and transportation of perishables.





PAGE 1

1 EXPLORATORY STUDY OF RFID APPLICATIONS FO R AIR CARGO OPERATIO NS By MAGALIE LANIEL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Magalie Laniel

PAGE 3

3 To all who guided me throughout my lifetime, making this milestone possible

PAGE 4

4 ACKNOWLEDGMENTS This work has been carried out under the guidance of my advisor Dr. Jean Pierr e mond. His brilliant ideas and suggestions more than helped shape this research. I will never be thankful enough for the opportunity and support that he gave me through all my studies I feel very privileged and fort un ate to have him as my advisor and wa nt to thank him for believing in me I would also like to thank each of my committee members, Dr. Ray Bucklin, Dr. Tom Burks, Dr. Daniel Engels and Dr. David Mikolaitis for their advice and contribution to my education thus far and in the future. Besides, I would like to express my thankfulness to Ismail Uysal who helped me extensively structure this dissertation and with whom I hope to have the pleasure to work with in the future. I would like to send a special thank to Trevor Howard and all of Air Canada Cargo for their generous implication and participation to this research. Without their help, many experiments would have simply been impossible to achieve. Moreover, I want to send a warm recognition to everyone at Franwell for helping in any way they cou ld and providing key equipment at critical times. Thank you for your trust and encouragement. I also want to thank Mr. Chris Noel from Sealed Air Corporation for providing temperature sensors in a timely manner and making my last experiment possible. I wou ld like to extend my gratitude towards my parents for their unconditional love and support regardless of distance and sacrifices I would also like to thank my other half Martin for his patience guidance and never ending re visions Last but not least I wo uld like to thank all my friends and colleagues for their support Particularly Cecilia, it was such a pleasure to shar e this path together

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBREVIATIONS ................................ ................................ ........................... 15 ABSTRACT ................................ ................................ ................................ ................... 16 CHAPTER 1 GENERAL INTRODUCTION ................................ ................................ .................. 18 2 RFID IN AIR CARGO: A LITTERATURE REVIEW ................................ ................. 24 Introduction ................................ ................................ ................................ ............. 24 RFID Technology and Definitions ................................ ................................ ........... 24 Radio Waves ................................ ................................ ................................ .... 25 Polarization of electromagnetic waves ................................ ....................... 26 Electromagnetic waves properties ................................ ............................. 27 RFID System Overview ................................ ................................ .................... 30 Readers ................................ ................................ ................................ ..... 30 Antenna ................................ ................................ ................................ ..... 31 Tags ................................ ................................ ................................ ........... 33 Frequencies ................................ ................................ ............................... 36 Air Cargo ................................ ................................ ................................ ................. 38 Aviation History ................................ ................................ ................................ 38 Air Cargo Supply Chai n ................................ ................................ .................... 39 Air cargo warehouse operations ................................ ................................ 40 Unit load device (ULD) ................................ ................................ ............... 41 Market ................................ ................................ ................................ .............. 42 Materials in Commercial Aircrafts ................................ ................................ ..... 43 Composites ................................ ................................ ................................ 43 Metals ................................ ................................ ................................ ........ 44 Electrical Systems in Commercial Aircrafts ................................ ...................... 45 Temperature Profile in Commercial Aircraft ................................ ...................... 46 Aircraft Safety ................................ ................................ ................................ ... 46 Cargo security and monitoring ................................ ................................ ... 46 Fire detection ................................ ................................ ............................. 47 Technologies ................................ ................................ .............................. 48 RFID in Aviation ................................ ................................ ................................ ...... 49 ISM Frequency and Aviation RFID Considerations ................................ .......... 49

PAGE 6

6 Aviation Applications ................................ ................................ ........................ 51 Passenger baggage sortation ................................ ................................ .... 51 Verification / authentication ................................ ................................ ........ 52 Tracking and locating ................................ ................................ ................. 53 Cargo ................................ ................................ ................................ ......... 54 Cold chain ................................ ................................ ................................ .. 56 Wireless Interference ................................ ................................ ....................... 58 Electronic devices ................................ ................................ ...................... 60 RFID interference ................................ ................................ ....................... 60 RFID airworthiness policy ................................ ................................ .......... 61 Concluding Remarks ................................ ................................ ............................... 62 3 AIR CARGO WAREHOUSE ENVIRONMENT AND RF INTERFERENCE ............. 68 Introduction ................................ ................................ ................................ ............. 68 Material s and Methods ................................ ................................ ............................ 72 Montreal Warehouse ................................ ................................ ........................ 74 Toronto Warehouse ................................ ................................ .......................... 74 Results and Discussion ................................ ................................ ........................... 74 433MHz ................................ ................................ ................................ ............ 75 915MHz ................................ ................................ ................................ ............ 76 2.45GHz ................................ ................................ ................................ ........... 77 Conclusion ................................ ................................ ................................ .............. 7 8 4 RADIO FREQUENCY PROPAG ATION INSIDE THE CARGO HOLD OF A DC 10 AIRCRAFT ................................ ................................ ................................ ......... 86 Introduction ................................ ................................ ................................ ............. 86 Materials and Methods ................................ ................................ ............................ 89 Test 1: Propagation Study ................................ ................................ ................ 90 Data analysis ................................ ................................ ............................. 90 Statistical analysis ................................ ................................ ...................... 94 Test 2: Validation of Relation between Signal Strength and Tag Reads .......... 95 Data point comparison ................................ ................................ ............... 96 Results and Discussio n ................................ ................................ ........................... 96 Test 1: Propagation Study ................................ ................................ ................ 96 Attenuation ................................ ................................ ................................ 96 Signal strength ................................ ................................ ........................... 99 Test 2: Validation of Relation between Signal Strength and Tag Reads ........ 100 Conclusion ................................ ................................ ................................ ............ 102 5 RADIO FREQUENCY INTERACTIONS WITH AIR CARGO CONTAINER MATERIALS FOR REAL TIME MONITORING ................................ ..................... 115 Introduction ................................ ................................ ................................ ........... 115 Materials a nd Methods ................................ ................................ .......................... 118 Test 1 ................................ ................................ ................................ ............. 120

PAGE 7

7 Test 2 ................................ ................................ ................................ ............. 120 Test 3 ................................ ................................ ................................ ............. 121 Results and Discussion ................................ ................................ ......................... 122 Test 1 ................................ ................................ ................................ ............. 122 433MHz ................................ ................................ ................................ .... 122 915MHz ................................ ................................ ................................ .... 123 2.45GHz ................................ ................................ ................................ ... 123 Test 2 ................................ ................................ ................................ ............. 125 Test 3 ................................ ................................ ................................ ............. 125 Conclusion ................................ ................................ ................................ ............ 127 6 TEMPERATURE MAPPING INSIDE AIR CARGO CONTAINERS DURING AIRSIDE OPERATIONS ................................ ................................ ....................... 131 Introduction ................................ ................................ ................................ ........... 131 Materials and Methods ................................ ................................ .......................... 134 Results and Discussion ................................ ................................ ......................... 136 During Flight ................................ ................................ ................................ ... 136 Before and After Flight ................................ ................................ ................... 138 Conclusion ................................ ................................ ................................ ............ 139 7 G LOBAL TRACKING SYSTEM FOR AIR CARGO SUPPLY CHAIN ................... 145 Introduction ................................ ................................ ................................ ........... 145 Typical Air Cargo Warehouse Operations ................................ ............................. 146 Typical Air Cargo Ramp Operations ................................ ................................ ..... 148 Findings from this Study and Recommendations for RFID Tracking System Implementation ................................ ................................ ................................ .. 149 Passive and Active RFID Tags ................................ ................................ ....... 150 ULD Materials ................................ ................................ ................................ 152 Frequency ................................ ................................ ................................ ...... 152 Conclusion ................................ ................................ ................................ ............ 154 8 GENERAL CONCLUSION ................................ ................................ .................... 160 APPENDIX A DC 10 CARGO HOLD AND CARGO DOOR SPECS ................................ ........... 162 B RADIO FREQUENCY ATTENUATION SURFACE PLOTS ................................ .. 164 C STATISTICAL ANALYSIS RESULTS FOR DC 10 RADIO FREQUENCY PROPAGATION ................................ ................................ ................................ ... 173 D RADIO FREQUENCY SIGNAL STRENGTH PROPAGATION SURFACE PLOTS ................................ ................................ ................................ .................. 179 E UNIT LOAD DEVICES, TEMPERATURE GRAPHS AND POSITIONS ................ 188

PAGE 8

8 LIST OF REFERENCES ................................ ................................ ............................. 194 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 205

PAGE 9

9 LIST OF TABLES Table page 2 1 EPCglobal tag class structure ................................ ................................ ............ 63 2 2 Permissible field strengths for RFID systems in accordance with FCC Part 15 (FCC, 2008). ................................ ................................ ................................ ....... 63 2 3 Aircraft radio syste ms ................................ ................................ ......................... 63 3 1 Receiving antenna specifications ................................ ................................ ....... 80 3 2 Minimum and maximum interference readings (in dBm) for six positions and three fre quencies at the Montreal warehouse. ................................ .................... 81 3 3 Minimum and maximum interference readings (in dBm) for nine positions and three frequencies at the Toronto warehouse ................................ ..................... 81 4 1 Specifications of the three RF systems used. ................................ ................... 104 4 2 Calculated parameters for the attenuation equation (Eq. 4 5) and maximum allowed output power adjustmen t. ................................ ................................ .... 104 4 3 Averages and standard deviations of attenuation levels for each test. ............. 104 4 4 Statistical analysis results for the effe ct of antenna location. ............................ 105 4 5 Statistical analysis results for the effect of antenna polarization. ...................... 105 4 6 Signal strength data for each test, averaged per vertical slice, and total cargo hold (Avg). ................................ ................................ ................................ ........ 105 4 7 Table summarizing the recorde d and adjusted power levels and read rates for circular and linear antennas across th e 12 cross sectional planes. ............. 106 5 1 Specifications of the six antennas used. ................................ ........................... 128 5 2 Signal strength measurements (dBm) for contr ol (no sample), test 1. Receiver antenna positions are measured from the emitting antenna. ............. 128 5 3 Signal strength deviatio n, test 1. Receiver antenna positions are measured from the emitting a ntenna and sample materials are positioned at ............... 128 5 4 Signal stren gth measurements plus signal strength deviation between material samples and control at th ree frequencies for test 2 ........................... 129 5 5 Signal strength measurement test 3. Receiver antenna positions are measured from the emitting antenna. ................................ ............................... 129

PAGE 10

10 5 6 Signal strength d eviation, test 3. Receiver antenna positions are measured from the emitting antenna and sample materials are positioned at ............... 129 6 1 Routes, aircraft and ULD specs from Toronto (YYZ). ................................ ....... 140 6 2 Temperature comparison between heated and unheated cargo holds inside an Airbus 330. ................................ ................................ ................................ .. 140 7 1 Current processes as well as proposed RFID solutions for the air cargo supply chain (cargo acceptance part). ................................ .............................. 155 7 2 Current processes as well as proposed RFID solutions for the air cargo supply chain (cargo build up part). ................................ ................................ ... 156 7 3 Current processes as well as proposed RFID solutions for the ai r cargo supply chain (cargo to/from the ramp section). ................................ ................. 157 7 4 Current processes as well as proposed RFID solutions for the air cargo supply chain (cargo break down and storage section). ................................ .... 158 7 5 Current processes as well as proposed RFID solutions for the air cargo supply chain (cargo delivery part). ................................ ................................ .... 159 C 1 Effects of fr equency, antenna location and antenna polarization on attenuation levels of the complete dataset. ................................ ...................... 173 C 2 Effect of width on attenuation levels for each frequency, antenna location and polariza tion. ................................ ................................ ................................ ...... 174 C 3 Effect of height on attenuation levels for each frequency, antenna location and polarization. ................................ ................................ ............................... 175 C 4 Effect of depth on attenuation levels for each frequency, antenna location and polarization. ................................ ................................ ............................... 176

PAGE 11

11 LIST OF FIGURES Figure page 1 1 Brief overview of the air cargo sup ply chain. ................................ ...................... 22 1 2 Some problems associated to the air cargo warehouse and aircraft operations. ................................ ................................ ................................ .......... 22 1 3 Some questions associated wi th warehouse and aircraft air cargo operations. 23 2 1 Electromagnetic spectrum ................................ ................................ .................. 64 2 2 Different parts of a wave (Lahiri, 2006) ................................ ............................... 65 2 3 Example of an RFID system on conveyor belt. ................................ ................... 65 2 4 Wave propagation for linear and circular polarization ................................ ......... 65 2 5 Cargo warehouse floor plan and activity areas. ................................ .................. 66 2 6 International flight temperature profile with both high temperature excursions (during sto povers) and low temperature excursions (in flight) ............................ 67 2 7 Example of a GSM/GPS capable RFID system for real time data acquisition (Schmoetzer, 2005). ................................ ................................ ........................... 67 3 1 Montreal cargo warehouse facility floor plan and interference reading points (numbered 1 to 6). ................................ ................................ .............................. 82 3 2 Toronto cargo warehouse facility floor plan and interference re ading points (numbered 1 to 9). ................................ ................................ .............................. 83 3 3 Noise floor of spectrum analyzer at three different resolution bandwidths. ......... 83 3 4 Worst cas e scenario for signal interfer ence readings around 433MHz. Span: 10MHz, RBW: 10kHz, attenuation: 0dB, gain: 0dBi. ................................ ........... 8 4 3 5 Worst case scenario for signal interf erence readings around 915MHz Span : 50MHz, RBW: 10kHz, attenuation: 0dB, gain: 2.5dBi. ................................ ........ 84 3 6 Worst case scenario for signal interference readings arou nd 2450MHz Span: 50MHz, RBW: 10kHz, attenuation: 0dB, gain: 8dBi. ................................ ........... 85 4 1 Section of an aircraft fuselage (Airbus A380) ................................ ................... 106 4 2 DC 10 30F from Arrow Cargo. ................................ ................................ .......... 107

PAGE 12

12 4 3 Typical fuselage section of a DC 10 30F, lower cargo hold circled in blue (Boeing, 2010c) ................................ ................................ ................................ 107 4 4 Cargo hold dimensions and RF emitting antenna positions. ............................. 108 4 5 Data point positions in the 3x3 grid. Twelve 3x3 grids are measured long the length of the cargo hold, every meter. ................................ .............................. 108 4 6 Tag readability test configuration. Tyvek sheet with 29 RFID tags (circled) covering half of the cargo hold cross section ................................ .................... 109 4 7 Attenuation surface plots for each vertical slice of 3x3 data point at 433MHz, circular antenna and top end antenna position. ................................ ................ 109 4 8 Attenuation surface plots for each vertical slice of 3x3 data point at 433MHz, circular antenna and center ceiling antenna position. ................................ ....... 110 4 9 Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz, circular antenna and top end antenna position. ................................ ................ 110 4 10 Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz, linear antenna and top end antenna position. ................................ ................... 110 4 11 Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz, circular antenna and center ceiling antenna position. ................................ ....... 111 4 12 Attenuation surface plots for each vertical slice of 3x3 data point at 2.45GHz, circular antenna and top end antenna position. ................................ ................ 111 4 13 Attenuation surface plots for each vertical slice of 3x3 data point at 2.45GHz, linear antenna and top end antenna position. ................................ ................... 111 4 14 Attenuation surface plots for each vertical slice of 3x3 data point at 2.45GHz, circular antenna and center ceiling antenna position. ................................ ....... 112 4 15 Distribution (in per centage) of each frequency tested, for circular antenna only and two antenna locations. ................................ ................................ ....... 113 4 16 Comparison of the change in average power levels and tag read rates for both antennas through lin ear regression. ................................ ......................... 114 5 1 Diagram of the anec hoic chamber setup for test 1. F our receiv er antennas shown fo r illustrative purposes. O ne receiver antenna is used at a time ......... 130 5 2 Anechoic chamber set up for test 2. A) The sample material is surrounded by foam absorber and placed one wavelength from the emitting antenna. B) The receiver antenna is taped behind the material sample. ................................ .... 130 6 1 ULD types and their respective tag positions. ................................ .................. 140

PAGE 13

13 6 2 Temperatures recorded for top and bottom tags during flight (gate to gat e). Data is congregated by total flight time and type of aircraft.. ............................ 141 6 3 Graph of averaged top and bottom tag temperatures during flight (gate to gate) for both short flights (1 2h) to and f rom Montreal (YUL). ......................... 142 6 4 Graph of averaged top and bottom tag temperatures during flight (gate to gate) for both medium short flights (4 6h) to and from Vancouver (YVR). ........ 142 6 5 Graph of averaged top and bottom tag temperatures during flight (gate to gate) for all 7 8h flights to and from London (LHR) or Frankfurt (FRA). ............ 143 6 6 Graph of averaged top and bottom tag temperatures during flight (gate to gat e) for the longest flight (>9h) from London (LHR) to Vancouver (YVR). ...... 143 6 7 Temperature profiles of all tags for ULDs A KH 1817 to and from Montreal ..... 144 7 1 Overview of the air cargo operations where the circled steps represent suggested RFID reading points. ................................ ................................ ....... 159 A 1 Standard cargo compartment and containers for model DC 10 series 10, 10CF, 30, 30CF, 40 and 40CF. ................................ ................................ ........ 162 A 2 Forward cargo loading door, model DC 10 series 10, 10CF, 30, 30CF, 40 and 40CF. ................................ ................................ ................................ ......... 162 B 1 Attenuation surface plot example one slice of data (dBm) ............................. 164 B 2 Attenuation surface plot for 433M Hz, top end antenna position and circular polarization. ................................ ................................ ................................ ...... 165 B 3 Attenuation surface plot for 433MHz, center ceiling antenna position and circular polarization. ................................ ................................ .......................... 166 B 4 Attenuation surface plot for 915MHz, top end antenna position and circular polarization. ................................ ................................ ................................ ...... 167 B 5 Attenuation surface plot for 915MHz, top end antenna position and linear polarization. ................................ ................................ ................................ ...... 168 B 6 Attenuation surface plot for 915MHz, center ceiling antenna position and circular polarization. ................................ ................................ .......................... 169 B 7 Attenuation surface plot for 2.45GHz, top end antenna position and circular polarization. ................................ ................................ ................................ ...... 170 B 8 Attenuation surface plot for 2.45GHz, top end antenna position and linear polarization. ................................ ................................ ................................ ...... 171

PAGE 14

14 B 9 Attenuation surface plot for 2.45GHz, center ceiling antenna position and circular polarization. ................................ ................................ .......................... 172 D 1 Signal strength surfa ce plot example one slice of data (dBm) ....................... 179 D 2 Signal strength surface plot for 433MHz, top end antenna position and circular polarization. ................................ ................................ .......................... 180 D 3 Signal strength surface plot for 433MHz, center ceiling antenna position and circular polarization. ................................ ................................ .......................... 181 D 4 Signal strength surface plot for 915MHz, top end antenna position an d circular polarization. ................................ ................................ .......................... 182 D 5 Signal strength surface plot for 915MHz, top end antenna position and linear polarization. ................................ ................................ ................................ ...... 183 D 6 Signal strength surface plot for 915MHz, center ceiling antenna position and circular polarization. ................................ ................................ .......................... 184 D 7 Signal strength surface plot for 2.45GHz, top end antenna position and circular polari zation. ................................ ................................ .......................... 185 D 8 Signal strength surface plot for 2.45GHz, top end antenna position and linear polarization. ................................ ................................ ................................ ...... 186 D 9 Signal strength surface plot for 2.45GHz, center ceiling antenna position and circular polarization. ................................ ................................ .......................... 187 E 1 Temperature profiles of all tags for ULDs (A) AKH 2084 and (B) AKH 977 8 to and from Montreal (YUL) ................................ ................................ .................. 188 E 2 Temperature profiles of all tags for ULDs AKH 1987 to and fro m Vancouver (YVR). ................................ ................................ ................................ .............. 189 E 3 Temperature profiles of all tags for ULDs (A) AKE 03782, (B) AKE 04090, (C) AKE 04969, and (D) AKE 05335 to and from London (LHR) ............................ 191 E 4 Temperature profiles of all tags for ULDs (A) AKE 03748, (B) AKE 04632, (C) AKE 05168, and (D) AKE 052 55 to and from Frankfurt (FRA). ........................ 193

PAGE 15

15 LIST OF ABBREVIATION S A320 Airbus 320 A321 Airbus 321 A330 Airbus 330 AKE a ir cargo container prefix for LD3 without forklift holes AKH a ir cargo container prefix for LD3 45 B 777 Boeing 777 dB d ecibel dBi d ecibel isotropic t he forward gain of an antenna compared with the hypothetical isotropic antenna, which uniformly distributes energy in all directions dBm d ecibel milliwatt p ower ratio in decibels (dB) of the measured po wer referenced to one milliwatt (mW ). FAA Federal Aviation Administration FCC Federal Communication Commission FRA Frankfurt international airport IC i ntegrated c ircuit LHR London Heathrow international airport RBW r esolution bandwidth RF r adio f requency RFID r adio f requency i dentification SNR signal to noise ratio ULD u nit l oad d evice YUL Montreal Trudeau international airport YVR Vancouver international airport YYZ Toronto Pearson international airport

PAGE 16

16 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EXPLORATORY STUDY OF RFID APPLICATIONS FO R AIR CARGO OPERATIO NS By Magalie Laniel August 2010 Chair: Ray A. Bucklin Ma jor: Agricultural and Biological Engineering The air cargo system is a complex network that handles a vast a mount of freight aboard passenger and all cargo aircraft. need for fresh products to be delivered y ear round all over the world, thus, temperature sensitive items are likely to be shipped by air because of their relatively short shelf life. Moreover, increasing demand for just in time delivery; container s being packed by third parties; inspection time b eing very limited and transportation security being linked to volume, monitoring of goods within containers as well as within the cargo warehouse may c ontribute to enhanced security, operation efficiency and provide valuable real time information. New t ech nologies to better track cargo shipments are accountable for maintain ing control and tracking along the supply chain. Radio frequency identification ( RFID ) is seen as an emerging technology for improving the air cargo supply chain For RFID technology to be implemented, more research has to be done regarding the environmental compatibility of air cargo warehouses, the regulations involved and the materials encountered in this supply chain. Moreover, the frequency of choice may be critical for system optimi zation. Therefore, the main objectives of this dissertation are: indentify the multiple RF interferences encountered inside an air cargo warehouse;

PAGE 17

17 evaluate the RF propagation behavior inside the cargo hold of a n aircraft at different frequencies ; v erify t he effect of container wall materials on RF propagation; study the temperature distribution of different cargo holds during flight. The main findings of this research are that interferences are l owest at 915MHz inside the air cargo warehouses studied. Foll owing the same direction, RF propagation inside the cargo hold was found to be best at 915MHz when taking into account federal spectrum regulations. In addition, the container materials experiment showed a very strong effect of aluminum on RF transmission and minimal interaction for all other composite materials. Moreover, there can be a significant temperature gradient between the top and bottom of air cargo containers during ground operations as well as during flights. The global system proposed from this research states that a combination of active and passive tags at 915MHz could create a well suited structure for tracking of the air cargo supply chain. To summarize, the findings of this dissertation suggest that using 915 MHz RFID systems for air cargo o perations would lead to the most success and system flexibility considering warehouse interference cargo hold RF propagation temperature monitoring needs and types of tag technology available today.

PAGE 18

18 CHAPTER 1 GENERAL INTRODUCTION Automatic identification (Auto ID) of objects enables the organizations that manage global supply chains to operate more efficiently and save cost. Auto ID includes a host of techno logies such as bar codes, smart cards, voice recognition, biometric technologies and radio frequenc y identification (RFID). Bar codes have been the primary means of identifying products since late 1960s. RFID offers many compelling advantages over bar codes, includ ing non line of sight operation, unique identifier, higher read rate volumes and sensor ca pabilities, to name a few. In addition, RFID technology enables computers to collect the unique ID assigned to item s. In combination with the Internet and associated infrastructure, RFID also allow s companies to track and trace individual items through th e supply chain. RFID aims to provide user s a near perfect supply chain visibility. That is, companies would be able to know exactly where every item in their supply chain is at any moment in time. The air carg o system consists of a large distribution netw ork linking manufacturers and shippers to freight forwarders to airport sorting and cargo handling facilities where shipments are loaded and unloaded from aircraft (Figure 1 1) Business and consumer demand for fast and efficient shipment of goods has fuel ed the rapid growth of the air cargo industry over the past 25 years. World air cargo traffic is forecasted to expand at an average annual rate of 5.8% for the next two decades, tripling current traffic levels (Boeing, 2008). The air cargo supply chain has been looking at RFID as a solution to increase its safety, operation efficiency and monitoring capability for many years (Figure 1 storing, building and breaking down shipmen ts, etc.) are not keeping up with the

PAGE 19

19 growing demand for fast and reliable shipping services. In addition, no time stamp is provided each time a shipment is loaded or unloaded from an aircraft, and only manual inspection tells if the shipment is in the rig ht aircraft or not. In air cargo operations, shipments are still being lost and items sometimes travel without their associated documents, which leads to claims that the carrier has to pay. Moreover, real time locating of loose goods as well as unit load d evices (ULD) in and out of the air cargo warehouse can provide visibility that not only the shipping company could benefit from; but also customers see value in knowing where their shipments are. The tracking and rapid locating of baggage, loose freight a nd containers (especially the associated integrity assurance of those items) is also essential to the overall security of a commercial flight. This tracking / locating of goods is accomplished today, for the most part, only by a very labor intensive manual process. The introduction of RFID to provide this asset tracking and locating offers the opportunity for: centralized monitoring; continuous surveying; automatic event logging; and, of course, more rapid finding of items when retrieval is mandatory (Cerin o and Walsh, 2000). When time becomes a primary consideration for delivery, air transportation is the mode of choice According to McCarthy (2003), o ne of the key drivers for the use of air cargo over other modes is the weight to value ratio of shipped goo ds Some specific market segments include : extremely high value products such as jewelry, luxury automobiles, and race horses; just in time products suc h as electronics and auto parts ; perishables such as fresh foods, flowers, and seasonal apparel ; and tim e sensitive products such as medical supplies and p harmaceuticals Many pharmaceutical and biotech products have a correlated sensitivity to temperature and high value (Wright,

PAGE 20

20 2008) which makes these industries a major customer of the air cargo industry. lives; therefore, temperature management of such shipment is of prime importance. delivered year round all over the world and with the cold supply chain requiring fast delivery; more and more perishable items are being shipped by air (Vega, 2008). Unfortunately, a faster transit time does not always imply controlled temperatu re throughout transp ortation. Of approximately 2.6 million tons of perishables air freighted in 2008, nearly 30% was estimated to be lost due to handling and temperature abuse (Catto Smith, 2006). RFID technology can also be combined with many dif ferent sensor applications, s uch as monitoring temperature, humidity, motion, etc. These features, with real time tracking of unique IDs throughout the air cargo supply chain open numerous valuable opportunities for shippers and customers. In essence, RFID is revolutionizing the way p roducts and goods are tracked and traced in the supply chain. It has been shown that RFID can significantly improve warehouse operation efficiency and supply chain performance (Chow et al., 2006; Poon et al., 2009; Vronneau and Roy, 2009; Visich et al., 2009; Wang et al., 2010). It has also been shown that RFID can improve the overall quality and shelf life of perishables through the cold supply chain (mond, 2007; Jedermann et al., 2007, 2009; Ruiz Garcia et al., 2008 ; Abad et al., 2009). Although, to im plement such technology in the air cargo industry, more research has t o be done regarding many compatibility aspects of RFID technology and the air cargo world.

PAGE 21

21 Commercial RFID systems are available under different standards, which work at different frequ encies. Choosing the best suitable frequency for an application depends on many factors. For instance, the read range needed, the size of tags preferred and the type of environment surrounding th e RFID system (materials and other interferences). In this re search, th ree frequenc ies will be ev aluated: 433MHz, 915MHz and 2.45GHz. Those frequencies are thought to be the most appropriate for the air cargo world today, m ostly because of their longer read ranges Objectives The main goal of this work is to eva l ua te the possibilities of using RFID to improve air cargo operations in general, as well as for perishable transportation. More specifically, some questions associated with air cargo operations will be addressed (Figure 1 3). Therefore, the four main objecti ve s of this dissertation are: Measure and evaluate the interference level at 433MHz, 915MHz and 2.45GHz in air cargo warehouses (chapter 3). Obtain 3D mapping of RF propagation inside a car go hold at 433MHz, 915MHz and 2.45GHz and compare with RFID tag rea dability at 915MHz (chapter 4) Evaluate RFID behavior around five air cargo container (ULD) materials at 433MHz, 915MHz and 2.45GHz (chapter 5) Study the temperature distribution inside air cargo containers in different cargo holds and aircrafts during f light (chapter 6)

PAGE 22

22 Figure 1 1. Brief overview of the air cargo supply chain Figure 1 2. Some problem s associated to the air cargo warehouse and aircraft operations.

PAGE 23

23 Figure 1 3. Some questions associated with warehouse and aircraft air cargo op erations.

PAGE 24

24 CHAPTER 2 RFID IN AIR CARGO: A LITT ERATURE REVIEW Introduction This review will give a general idea o f radio frequency identification ( RFID ) technology, its definitions and the way systems work. Then will follow an overview of the air cargo wor ld, aviation history, the current market status, plus a brief description of the major parts and their operations. In addition, aircraft construction, systems, avionics and safety issues will be discussed. Subsequently, the subject of RFID in aviation will be described through regulations, applications as well as possible interference of the technology. The conclu ding remarks comment on the potential of RFID to improve air cargo operations in general. RFID Technology and Definitions Radio frequency identifi cation (RFID ) is an automatic wireless data collection technology with a long history. The fundamentals of RFID technology are based on and radio waves as forms of elec tromagnetic energy back in 1846. For the last two decades RFID tags have been used in many applications (e.g. automatic toll roads, smart cards, store theft protection, access control, animal tracking, item tracking, etc.) which supply chain management and item tracking have been the fastest growing areas (Landt, 2005). Improvements in semiconductor technology resulted in reduction in the size of circuitry, reduction in cost of tags, increased functionality, and increased reliability which sped up the indu strial applications of RFID (Landt, 2005).

PAGE 25

25 Radio Waves Radio waves account for a portion of the electromagnetic spectrum ( Figure 2 1). Radio waves at their most basic are considered as wave forms of electrical and magnetic fields and as a result, have ampl itude, wavelength ( ), velocity ( v ) and frequency ( f ), the relationship of which is expressed as: (2 1) Electromagnetic waves are created by electrons in motion and consist of oscillating electric and magnetic fields. These waves can pass through a num ber of different material types (Lahiri, 2006) The highest point of a wave is called a crest and the lowest point is called a trough as shown in Figure 2 2 The distance between two consecutive crests or two consecutive troughs is called the wavelength One complete wavelength of oscillation of a wave is called a cycle The time taken by a wave to complete one cycle is called its period of oscillation The number of cycles in a second is called the frequency of the wave. The frequency of a wave is measu red in hertz (abbreviated as Hz) and named in honor of the German physicist Heinrich Rudolf Hertz. If the frequency of a wave is 1 Hz, it means that the wave is oscillating at the rate of one cycle per second. It is common to express frequency in KHz (or ki lohertz = 1,000 Hz ), MHz (or megahertz = 1,000,000 Hz ), or GHz (or gigahertz = 1,000,000,000 Hz ). Amplitude is the height of a crest or the depth of a trough from the undisturbed position (Lahiri, 2006) Radio waves can be further divided up into groups; Low Frequency (LF), High Frequency (HF), Ultra High Frequency (UHF) and Microwave Frequency (MF) with similar categories applying to RFID systems. Electromagnetic energy has been best

PAGE 26

26 described as a stream of photons each traveling at the speed of light in a wave like pattern. An electromagnetic wave propagates in a direction that is at right angles to the vibrations of both the electrical and magnetic oscillating fields ( Winder and Carr 2002). Radio waves and microwaves are situated towards the lower end of the electromagnetic spectrum (Figure 2 1) meaning that waves situated in the low frequency category posses lower amounts of energy ( E ) compared to microwave frequency waves (2 2) onstant. Electromagnetic waves can be characterized in terms of frequency, wavelength or 8 ) as the velocity, it is now possible to say (Meyers et al., 2007): (2 3) Polarization of electromagnetic waves The polarization of an electromagnetic wave is determined by the direction of the electric field of the wave. There is a differen ce between linear polarization and circular polarization. In linear polarizat ion the direction of the field lines of the electric field in relation to the surface of the earth provide the distinction between horizontal (the electric field lines run ning parallel to the surface of the earth) and vertical (the electric field lines run ning at right angles to the surface of the earth) polarization ( Finkenzeller, 2003)

PAGE 27

27 The transmission of energy between two linear polarized antennas is optimal if the two antennas have the same polarization direction. Energy transmission is at its lowest point, on the other hand, when the polarization directions of transmission and receiving antennas are arranged at exactly 90 or 270 in relation to one another (e.g. a horizontal antenna and a vertical antenna). On the other hand, circular polarization oc curs when the polarization direction of the electromagnetic field generated rotates through 360 every time the wave front moves forward by a wavelength. The rotation direction of the field can be determined by the arrangement of the delay line. We differe ntiate between left handed and right handed circular polarization ( Finkenzeller, 2003). Electromagnetic waves properties In free space, all electromagnetic waves obey the inverse square law which states that the power density of an electromagnetic wave is proportional to the inverse of the square of the distance from the source In other words, as the separation distance is doubled, the electrostatic force is decreased by a factor of four ( Henderson, 2010 ) As the electromagnetic waves propagate in their en vironment, they encounter many objects and behave differently around those obstacles, sometimes in a critical way towards the communication link needed for the functioning of an RFID system. It has been reported that environmental factors may decrease the reader range of passive RFID systems by at least 50% (Keskilammi et al., 2003). It is also well established that higher frequencies experience greater attenuation levels than lower frequencies (Keskilammi et al., 2003). Any wave incident upon an object wil l penetrate the material, a portion may be transmitted and another portion may also be reflected back into the environment. The exact amount of transmission and reflection is also dependant on the angle of incidence, material thickness, and dielectric prop erties (Blaunstein and

PAGE 28

28 Christodoulou, 2007). Part of the high frequency energy that reaches the object is absorbed by the object and converted into heat; the rest is scattered in many directions with varying intensity (Finkenzeller, 2003). Radio waves can be affected by the material through which they propagate. A material is called RF lucent for a certain frequency if it lets radio waves at this frequency pass through it without any substantial loss of energy. A material is called RF opaque if it blocks, reflects, and scatters RF waves. A material can allow the radio waves to propagate through it but with substantial loss of energy. These types of materials are referred to as RF absorbent. The RF absorbent or RF opaque property of a material is relative, b ecause it depends on the frequency. That is, a material that is RF opaque at a certain frequency could be RF lucent at a different frequency (Lahiri, 2006). The presence of more than one wave in a space may result in interference between the waves, which can be constructive (they reinforce one another), or destructive (cancel each other in whole or in part). There are a number of different ways an electromagnetic wave may interact with materials in its surrounding area as follows (Wu et al., 2006; Domdouzi s et al., 2007): Scattering : This occurs when a wave hits an obstacle smaller than its wavelength. It leads to the formation of scattered waves which are redirected with random phase and amplitude (Blaunstein and Christodoulou, 2007). This can be the resul t of rough surfaces, small objects or irregularities in the transmission medium. Refraction : This is the change in direction of a wave due to a change in its speed. This is most commonly observed when a wave passes from one medium (with a certain

PAGE 29

29 refractio n index) to another at an angle Refraction is described by Snell's law, which states that the angle of incidence is related to the angle of refraction (Reed, 2009) Fading : This is a variation of the signal strength with time ( Meyers et al., 2007). It o ccurs due to time dependent changes in multipath. Fade zones are small areas inside the interrogation zone that lead to periodic attenuation of the received signal. This effect increases with the distance from the emitting antenna. This occurrence is too r andom to make possible the prediction of signal strength at a particular point in time ( Mac Carthy 2009). Multipath : This occurs when a radio wave arrives at a particular receiving antenna from more than one propagation route due to its interactions with the surrounding environment (Lahiri, 2006). Multipath, or path loss strongly depends on propagation environment. Reflection and cancellation : The electromagnetic field emitted by the reader is not only reflected by a transponder, but also by all objects i n the vicinity, the spatial dimension s of which are greater than the wavelength of the field (Rappaport, 2002) The reflected fields are superimposed upon the primary field emitted by the reader. This leads alternately to a local damping or even so called cancellation (anti phase superposition) and amplification (in between the individual minima. The simultaneous occurrence of many individual reflections of varying intensity at different distances from the reader leads to a very erratic path of field strength around the reader, with many local zones of cancellation of the field. Such effects should be expected particularly in an environment containing large metal objects ( Finkenzeller, 2003) The importa nce of these properties cannot be

PAGE 30

30 over emphasized as they are all hugely important in relation to passive UHF RFID systems as they impact on how the electromagnetic wave (essential for coupling) is affected by different objects. Direct penetration: Little or no reflection occurs when electromagnetic waves penetrate directly through objects such as paper, non conductive plastics or textiles (Penttil et al., 2006). These materials, including most composites, are non absorbing and possess low refractive index es. Such materials are generally referred to as being RF lucent. Diffraction: Similarly to light propagation, materials surrounding radio waves can provoke diffraction of the waves, which can lead to RF signal variations. It is described as the apparent b ending of waves around small obstacles the spreading out of waves past small openings or t he deviation in the path of a wave that encounters the edge of an obstacle At high frequencies, diffraction, like reflection, depends on the geometry of the object, as well as the amplitude, phase, and polarization of the incident wave at the point of diffraction ( Rappaport, 2002) RFID System Overview A basic RFID system consists of a computer with software connected to a reader and one or more reader antennas, which communicate wirelessly with tags (Figure 2 3). The reader transmits an RF signal to the tags via its antenna (s) The tags receive power from the antenna and then send their information back Following is a description of each component in more detail. Rea ders RFID readers can be portable or fixed, depending on the application. A typical RFID reader has both transmitting and receiving functions for data transfer and

PAGE 31

31 communication with tags (Keskilammi et al., 2003 ; Poussos and Kostakos, 2009). Most interrog ators consist of an RF transceiver module (transmit and receive), a signal processor, a controlling unit and a coupling element (antenna) and data interface to a host system (Lahiri, 2006). EIRP: The E quivalent I sotropic R adiated P ower (EIRP) determines th e power of the signal transmitted by the reader in the direction of the tag. Maximum allowed EIRP is limited by national regulations (e.g. in North America it is 4W) (Nikitin and Rao, 2006). Reader sensitivity: is another important parameter which defines the minimum level of the tag signal which the reader can detect and resolve. The sensitivity is usually defined with respect to a certain signal to noise ratio or error probability at the receiver. Factors which can affect reader sensitivity include receiv er implementation details, communication protocol specifics, and interference, including signals from other readers and tags. An ideal reader can always detect an RFID tag as long as the tag receives enough power to turn on and backscatter (Nikitin and Rao 2006). Antenna Antennas are essential components of both RFID tags and readers. Their principal function is the facilitation of a dual directional communication link between tag and reader ( Dobkin, 2008 ). At its most basic an antenna is a particular arra ngement of conductors designed to transmit an electromagnetic field in response to the application of an alternating electric current. It also has the ability to generate a voltage between terminals when placed in a time varying electromagnetic field (Fink enzeller, 2003). Both RFID tags and reader antennas come in a variety of sizes and shapes which determines their operational characteristics (polarization for example).

PAGE 32

32 Reader a ntenna : These are used to communicate with the nearby tags. Antennas have emit ting and receiving capabilities. The antenna first propagates the RF wave into the environment in order to establish a communication link between the tags and reader to facilitate coupling (Liang et al., 2006). This RF wave creates an interrogation zone wh ich is an area surrounding the antenna where communication will take place provided that a tag is present. Once communication i s established, the antenna also receive s a resultant signal from the tag which is transferred to the reader for demodulation. Th e antenna may be physically incorporated into the reader, which is generally the case for handheld units. Alternatively, the antenna might be individually housed and attached to the reader by appropriate cables (Roussos and Kostakos, 2009 ). Depending on th e desired use, readers can support connection to more than one antenna at a time. Tag antenna : They operate on the same principle as reader antennas, but face some different practical challenges (Dobkin, 2008). In the case of passive RFID tags, the tag an tenna is responsible for receiving the electromagnetic wave from the reader antenna and reflecting the modulated backscatter signal to the reader. For active RFID tags, the antenna is responsible for emitting the internally generated signal. The size and s hape of the antenna determine s the operating frequency as well as the application and ideal orientation of the tag (Fuschini et al., 2008). Tag antenna must also be small enough to fit the item it is identifying; have omnidirectional and hemispherical cove rage; have a polarization that matches with the reader; and be robust and low cost (Keskilammi et al., 2003). Polarization : Propagating electric fields point in a certain direction in space. Polarization refers to the orientation of the electric field radi ated by the antenna. If the

PAGE 33

33 vector rotates with time, then the wave is elliptically polarized. The degree of ellepticity from a circle to a straight line gives circular and linear polarization. For linear polarization (vertical / horizontal), the vector os cillate on one plane as the wave propagates whereas in circular polarization it rotates through 360 degrees per cycle ( Figure 2 4). Gain : The gain, or amplification factor, is the factor by which the input power is amplified Ideally a reader antenna has a high gain due to the fact that the received power at the tag is directly proportional to the reader antenna gain. Tags An RFID tag is a device that can store and transmit data to a reader in a contactless manner using radio waves. Tags have three main c omponents: an integrated circuit (IC or chip ), an antenna and a substrate (Meyers et al. 2007 ). They are available in a wide variety of sizes, shapes and functionality which determines their unit cost (Jedermann et al., 2009). The tag is responsible for s toring and sharing user defined information regarding the item to which it is attached. It may be constantly in an active state whereby it may record storage conditions in its immediate environment or it may remain in a dormant state until activated by an interrogating wave from a nearby reader depending on its classification (Nikitin and Rao, 2006). It communicates with the reader by superimposing its stored data (through signal coding and demodulation). Transponders can be classified according to sources of energy : Passive tags Active tags Semi active tags Semi passive tags

PAGE 34

34 Passive t ags : Passive RFID tag s have no internal power source (no battery). In inductively coupled systems, when the tags are present in the RF field of an RFID interrogator, the ene rgy induced on the tag circuitry is used for transmitting back the ID of the tag. In UHF systems, electromagnetic backscatter coupling is used at the tag circuitry for changing the impedance of the tag antenna according to its ID. A passive tag is simple i n its construction and has no moving parts. As a result, such a tag has a long life and is generally resistant to harsh environmental conditions. Passive tags can be very small and low cost to manufacture. On the other hand, they have limited data capacity and shorter read range. In tag to reader communication for this type of tag, a reader always communicates first, followed by the tag. The presence of a reader is mandatory for such a tag to transmit its data (Lahiri, 2006). Active t ags : Active tags are be aconing in a defined period of time by using power for data transmission. The on board electronics can contain microprocessors, sensors (temperature, humidity, motion, e tc.), and input / output ports powered by the on board power source. Therefore, for example, these components can measure the surrounding temperature and generate the average temperature data (Keskilammi et al., 2003). The components can then use this data to determine other parameters such as the expiry date of the attached item. The tag can then transmit this information to a reader. In tag to reader communication for this type of tag, a tag always communicates first, followed by the reader (Lahiri, 2006) Active tags also have longer read ranges than passive tags. They are ideal in environments of high electromagnetic interference because of their ability to broadcast a stronger signal with the aid of their internal power

PAGE 35

35 source (Jeddermann et al., 2009). Some advantages and disadvantages of active tags are: Increased functionality (sensor, monitoring, recording) Long read ranges Large memory capacity Physically bulky High production costs Fragile because of moving parts in the design Life span limited to power source These tags are commonly used in RTLS (Real Time Location Systems) in which the tag continuously reports its ID to the receiver units and location of the tag is usually calculated by using RSS (Received Signal Strength) information and triangu lating between different receivers ( Altunbas, 2010 ). Semi active tags : The name associated with this type of tag is not yet widely accepted. It is somewhat a type of active tag that enters a sleep or a low power state in the absence of interrogation by a r state by issuing an appropriate command. This state saves the battery power, and therefore, a tag of this type generally has a longer life compared to an active transmitter tag. In addition, because the tag transmits only when interrogated, the amount of induced RF noise in its environment is reduced. Semi passive t ags : These combine both passive and active type technology. Semi passive systems have internal batteries but are not beaconing signals in a de fined period. In the presence of the RF field of an interrogator, the tag wakes up and starts to backscatter its ID to the interrogator (like a passive tag); but using its integrated battery supply to increase the signal (McCarty, 2009). The trade off for energy efficiency is to

PAGE 36

36 have reduced response time caused by the time slot needed to wake up the transponder. Tag generation and classification Tags can be classified according to their power source, as seen earlier, but they may also be grouped according to their functionality (Meyers et al., 2007) This classification is based on EPCglobal standard as shown in Table 2 1 This classification of ta g is based on the format, read / write capability and programming capability. The EPC classification consists of Class and Generation. The Class describes a tags basic functionality, for example whether it has memory or an on board power source, whereas Generation refers to a tag specification's major release or version number (Khan et al., 2009). Frequencies Comm ercial RFID systems designed for different applications can work at different frequencies, such as: Low frequency (LF): Frequencies between 30 kHz and 300 kHz are considered low, and RFID systems commonly use the 125 kHz to 134 kHz frequency range. At LF, the power supply to the transponder is generated by inductive coupling, which means short read ranges between the tags and reader antenna. RFID systems operating at LF generally have low data transfer rates from the tag to the reader, and are especially good i f the operating environment contains metals or liquids (Lahiri, 2006), which is why LF systems are commonly used in animal identification applications. High frequency (HF): HF ranges f rom 3MHz to 30MHz, with 13.56 MHz being the typical frequency used for H F RFID systems. A typical HF RFID system uses passive tags, has a slow data transfer rate from the tag to the reader, and offers fair performance in the presence of metals and liquids (Lahiri, 2006). Moreover, HF systems

PAGE 37

37 also work within short read ranges (generally within 1m), it is used in smart cards applications like access control or contactless payments. Ultra high frequency (UHF): UHF is the most common passive RFID tags used in supply chain applications worldwide. While the entire UHF spectrum range s from 300 MHz to 1 GHz it operates between 902 928MHz in the Americas; 865 868MHz in Europe, middle east and Russia; and 866 869MHz and 923 925MHz in Asia, Australia and the Pacific. The accepted standard for passive UHF frequency is ISO 18000 6C (UHF Gen2 ). Active or semi active RFID systems in UHF frequencies operate at 433MHz. ISO 18000 7 is the accepted standard for parameters of active air interface communications at 433MHz. A UHF system can therefore use both active and passive tags and has a fast dat a transfer rate between the tag and the reader, but performs poorly in the presence of metals and liquids ( not true, however, in the cases of low UHF frequencies such as 433MHz) (Lahiri, 2006). UHF systems work ing in the electromagnetic field offer a much longer read range than lower frequencies. Microwave frequency: Microwave frequenc ies range upward from 1GHz. A typical microwave operating frequency is either at 2.45GHz or 5.8GHz in the Industrial Scientific and Medical (ISM) band. Microwave systems can b e either passive or semi passive and provide the fastest data communication rates compared to the other frequencies (Lahiri, 2006). Microwave frequency, like UHF, offers long read ranges, especially when working in an active RFID system. Read range perform ance for passive systems in the presence of water and metallic surfaces is very poor because of the higher signal attenuation at higher frequency (Friis, 1946).

PAGE 38

38 National licensing regulations In the USA, RFID systems must be licensed in accord ance with li censing regulation FCC Part 15 from the Federal Communications Commission. This regulation covers the frequency range from 9kHz to above 64GHz and deals with the intentional generation of electromagnetic fields by low and minimum power transmitters (intent ional radiators) plus the unintentional generation of electromagnetic fields (spurious radiation) by electronic devices such as radio and television receivers or computer systems. The category of low power transmitters covers a wide variety of applications for example cordless telephones, biometry and telemetry transmitters, on campus radio stations, toy remote controls and door openers for cars. Inductively coupled or backscatter RFID systems are not explicitly mentioned in the FCC regulation, but they au tomatically fall under its scope due to their transmission frequencies, which are typically in the ISM bands, and their low transmission power (Finkenzeller, 2003). Table 2 2 l ists some of the frequency ranges that are important for RFID systems. Air Cargo Aviation History Times have changed since the Wright Brothers and the first flight of a powered aircraft in 1903 (Taylor 1989; Bilstein 1994; Wegener 1997). They were not only pioneers of flight, they also were the first ones to ship goods by air, when in November 1910, a department store from Ohio made arrangement with them to have a bolt of silk flown up from Dayton to Columbus (Bilstein, 1994). Not long after, in 1911, the first official airmail flight is made in India, where 6,500 letters were carried over about 1 0 km (Taylor, 1989) It is only after World War I on May 15, 1918 that air mail service began in the US between New York City, Philadelphia and Washington D.C. with a JN 4, which

PAGE 39

39 was built as an Army training airplane. In 1919, American Railway Express made an unsuccessful attempt to deliver cargo to Chicago (Bilstein, 1994). In the mean time, the Airmail Act (1925) and the Air Commerce Act (1926) were created. Later, in 1927, plans were made for four commercial airlines (National Air Transport, Colonial Air transport, Boeing Air Transport and Western Air Express) to fly express. From hesitant beginnings to slow progress, commercial air cargo made great strides in the post World War II era. Wartime experience in long range cargo operations helped but the postwar availability of dozens of surplus military multiengine transports was more important Unfortunately, the maintenance cost of those aircraft only allowed a handful of companies to survive. On the other hand, the scheduled passenger lines, sensing lost revenues, began to pay more attention to cargo services in their normal pa ssenger routes and began to operate their own all cargo services (Bilstein, 1994). Air Cargo Supply Chain The air cargo system consists of a large, complex distribution network linking manufacturers and shippers to freight forwarders to airport sorting and cargo handling facilities where shipments are loaded and unloaded from aircraft ( Elias, 2007 ) The airport forms an essential part of the air cargo supply chain, because it is the physical site at which a modal transfer of transport is made from the air m ode to land mode. It is the point of interaction between the airline and the user (Ashford et al., 1983). Airports are divided into landside (parking lot, access roads, etc.) a nd airside (all areas accessible to aircraft, including runways taxiways and ramps ) areas. In addition to people, airports move cargo a round the clock. Cargo airlines often have their own on site and adjacent infrastructure to transfer parcel s between ground and air.

PAGE 40

40 Air cargo warehouse operations Figure 2 5 prese nts the layout of an The cargo terminal is divided into an import area and an export area. The flow of goods through the terminal is either from the airside to the landside (terminating freights or connecting freights requirin g the road feed service), from the landside to the airside (originating freights or connecting freights arriving from a road feeder service), or from the airside to the airside via the terminal (connecting freights). Export area: The export area is dedicat ed to receiving, processing and preparing outbound freights, which refers to all shipment moving from an outside customer, and area, either as bulk or as shipper loa ded unit device. The freight gets weighed and dimensioned by the acceptance agent and stored at the appropriate location depending on its flying time and destination. If items are bulk, they ultimately go to the build up area to be put in a ULD (Unit Load Device) or are transported in a tub cart directly to the airplane if this airplane is bulk loaded. ULDs are transported onto roller system through the cargo facility and onto trailers to the airside. All export shipments leave the Import area: The import area is dedicated to receiving, processing and releasing inbound freights which refers to all shipments coming from a flight, going to an outside customer. ULDs are transported the same way between the airside and cargo facility (trailers). Bulk is unloaded from the aircraft directly into tub carts. Everything is brought down when needed and stored until customer pick up.

PAGE 41

41 The movement of transiting goods (from one flight to another flight) also go es before exiting the warehou Unit load device (ULD) Unit Load Devices (ULDs) play a vital part in ensuring that as air cargo volumes increase; they are moved safely, quickly and cost effectively (IATA, 20 0 2 ) ULD is the c orrect terminology used by the air transport industry for containers and loading units that are used for the carriage of cargo by air. It allows large quantities of cargo to be bundled into large units. Pallets and nettings as well as rigid containers are commonly used for freight transport by air. Each ULD is required to have a marking that identify its type code, maximum gross weight and actual tare weight (IATA, 20 02 ) Currently, technical specifications for unit load devices are set by the International Air Transport Association (IATA). While the world is talking about climate change, the airline industry is looking at ways to be more fuel efficient to minimize their operational costs as well as their impact on the environment. One w ay to do so is to red uce weight, minimizing weight without compromising the business volume is feasible by using lighter containers, or ULDs. Composite ULDs can save up to 25% of the tare weight of a traditional aluminum ULD (Nordisk, 2010). Kevlar ULDs are constantly replaci ng older aluminum containers and up to 43% of their fleet, whereas Lexan containers count for the remaining 18%.

PAGE 42

42 Market During the late 1960s, the total tone kilometerag e of freight doubled every four years, an average annual growth rate of 17%. At that time, the aviation world was replete with extremely optimistic forecasts of a burgeoning air cargo market. The prolonged and recurrent economic recessions and the tenfold increase in oil prices of the 1970s militated against sustained growth in North America (Ashford et al., 1983). World air cargo traffic grew 5.1 % in 2007, which followed 3.2 % growth in 2006 and 1. 7% growth in 2005, making those three years the weakest growth period for the industry since the first Gulf War, 1990 1992. Tepid traffic growth can be largely attributed to high fuel prices, which were increasing from late 2003 through July 2008. In response to the ongoing rise in jet fuel prices, freighter operators have accelerated fleet renewal activities, most notably in the large wide body sector. Freighters count for about 10% of the total airplane fleet ( Boeing, 2008). A wide body aircraft is a large airliner with two passenger aisles whereas a narrow body only has one passenger aisle. A few decades ago, it was hard to foresee the present degree of traffic volume in av iation. An increase of up to 5.8% per year is estimated for the next two decades which will mean triple the amount of the present cargo traffic volume in the next 20 years (Boeing, 2008). As a result of this increase, the focal point in aviation research h as changed: Socio economic aspects are coming to the fore. The reduction of emissions such as noise and pollutants is becoming more significant. In particular, the reduction of the weight of the structure of future aircraft is a central task that will enab le a reduction of fuel consumption and an increase in the payload (Wilmes et al., 2002).

PAGE 43

43 The international competitive situation and the related increase in global competition in the aircraft industry is additionally making it necessary to considerably red uce costs in the development, production, and maintenance of the next generation of aircraft. The development time must be considerably shortened in order to enter aircraft faster into service. In addition to weight and cost, additional challenges in the f uture will be increased safety requirements for aircraft in the case of accidents, etc. Improvements in these areas are indispensable in order to ensure a high acceptance of this means of transportation in the future (Wilmes et al., 2002). Materials in Com mercial Aircrafts Commercial aircraft include types designed for scheduled and charter airline flights, carrying both passengers and cargo. The larger passenger carrying types are often referred to as airliners the largest of which are wide body aircraft Some of the smaller types are also used in general aviation Aircraft construction materials used to be mostly aluminum alloys, but nowadays more and more composites are utilized in aircrafts design. Composites The development of composite materials is c onsidered to be one of the most important advances in aviation design since aluminum was introduced in the 1920s. Development of various composite materials has had a very positive impact on the performance, shape, reliability, weight, cost and composition of modern aircraft Composites are a combination of two or more significantly different inorganic or organic components. Although the components together form a composite material they each maintain their original form and do not blend together. In a comp osi te material, one component serve s

PAGE 44

44 with the other component or components serving as reinforcement. An epoxy resin matrix with glass fiber reinforcing is one of the more commonly known co mposite materials, but continuing research is resulting in the production of various other composite materials which are proving beneficial in aviation design as well as in other industries (Anonym, 2007) Despite their strength, light weight, long life ex pectancy, corrosion resistance, and resistance to damage from cyclic loading (fatigue); composites have not been a miracle solution for aircraft structures. Composites are hard to inspect for flaws and can sometimes be brittle Some of them absorb moisture Most importantly, they can be expensive, primarily because they are labor intensive and often require complex and expensive fabrication machines. Aluminum, by contrast, is easy to manufacture and repair (Day, 2009) Modern airliners use significant amou nts of composites to achieve lighter weight. About 10% of the structural weight of the Boeing 777, for instance, is composite material (Day, 2009) The new Boeing 787 Dreamliner has made extensive use of composite materials, resulting in a lighter weight a irplane which is expected to have a number of benefits including greater fuel efficiency. This twin engine, wide body jet airliner is constructed from 50% composite with aluminum, titanium and steel making up 45% and a variety of components making up the b alance of 5%. The composite material most used in the Boeing (Boeing, 2010a) Metals Aluminum still remains a re markably useful material for aircraft structures and metallurgists have worked hard to develop better aluminum alloys (a mixture o f aluminum and other materials) Aluminum is a very tolerant material and can take a

PAGE 45

45 great deal of punishment before it fails. It can be dented or punctured and still hold together (Day, 2009) Aluminum alloys used in the aerospace industry are high strength and able to perform well in harsh and challenging environments. 7075 Aluminum is the alloy of choice when it comes to manufacturing aircraft parts and 5052 aluminum, which is not quite as strong but has more weldability, is sometimes used. 7075 c ontain s zinc and copper which is ideal for highly stres sed parts and is considered the strongest type of aluminum. It has good high temperature resistance and corrosion resistance, both necessary charac teristics in aircraft aluminum Aircraft metal must be strong yet lightweight at the same time, and aluminum exhibits a good strength to weight ratio, making it the first choice in airplane construction. The airframe of a typical commercial transport aircraft is 80 % aluminum by weight (The Aluminum Association, 2008) Electrical Systems in Commercial Aircrafts Ai rcraft power can be generated by DC or AC power sources. DC systems usually sources which are three phase systems where three sine waves are generated 120 degrees out of phase f rom each other. In this layout the phase voltage of a standard aircraft system is 115 VAC and the standard frequency is 400 Hz ; which is the same standard for ground power at most airports (Moir and Seabridge, 2001). Avionics : This is a portmanteau word of "aviation electronics". It comprises electronic systems for use on aircraft comprising communications navigation and the display and management of multiple systems. Table 2 3 lists the main systems and their respective working frequencies.

PAGE 46

46 Temperature Profile in Commercial Aircraft Temperat ure is well regulated in the cabin of most passenger flights, but it is not necessarily the case inside the cargo hold or of freighter flights. Temperature distribution around cargo depends on many factors, to name a few: weather, duration of flight, type of aircraft (ability to control cargo ambient temperature), altitude, transit time (on tarmac), etc. As shown in Figure 2 6, a study on an international shipment of live mice during summer showed that both heating problems (during airport handling) and coo ling problems (during flight) can occur ( Syversen et al., 2008) Aircraft Safety Every day approximately six million people board airplanes and arrive safely at their destinations. Flying is one of the safest modes of transportation today. The overall saf ety record of commercial airplanes is excellent and has been steadily improving over time. During the 1950s and 1960s, fatal accidents occurred about once every 200,000 flights. Today, the worldwide safety record is more than ten times better, with fatal a ccidents occurring less than once every 2 million flights (Boeing, 2010b) Cargo security and monitoring The air cargo system is a complex, multi faceted network that handles a vast a mount of freight, packages, and mail carried aboard passenger and all car go aircraft. The air cargo system is vulnerable to several security threats including potential plots to place explosives aboard aircraft; illegal shipments of hazardous materials; criminal activities such as smuggling and theft; and potential hijackings a nd sabotage by persons with access to aircraft. Several procedural and technology initiative s to enhance air cargo security and deter terrorist and criminal threats have been put in place or are under consideration. Procedural initiatives include industry wide

PAGE 47

47 increased physical security of air cargo facilities; increased oversight of air cargo operations; security training for cargo workers; and stricter controls over access to carg o aircraft and air cargo operations areas. Technology being considered to improve air cargo security includes tamper resistant and tamper evident packaging and containers; explosive detection systems (EDS) and other cargo screening technologies; blast resi stant cargo containers and aircraft hardening; and biometric systems for worker identification and access control (Elias, 2007) While the primary policy focus of legislation has been on cargo carried aboard passenger aircraft, air cargo security also pres ents a challenge for all cargo operators (FAA, 2006) History shows that very few accidents were caused by hazardous cargo content. According to the N ational T ransportation S afety B aviation accident database (NTSB, 2010), in the last 15 years, less than 20 accidents occurred from that cause, which corresponds to less than 0.1% of all accidents and incidents within that period of time. Moreover, only one major (fatal) accident was caused by fire in a cargo hold. Fortunately, the low numbers of accide nts do not slow down the aviation authorities in encouraging the design of safer aircrafts. Fire detection For some aircraft compartments a fire / smoke detection system is required by the regulations JAR (Joint Aviation Requirements) and/or FAA. In additi on aircraft manufacturers install supplementary fire / smoke detection systems to increase the level of safety. These systems must comply e.g. with the regulations ( Schmoetzer, 2001) The urgency of the corrective action subsequent to a fire / smoke warni ng depends directly on the risk and is reflected in the procedure s to be applied by cockpit or cabin

PAGE 48

48 crew. For example, a cargo compartment smoke warning is indicated to the flight deck crew as a red warning, this means the crew has to perform the action i mmediately As long as the crew is unable to differentiate between a true and a false warning, it has to follow the certified procedure. The impact of a false fire / smoke warning in non accessible compartments is extensive and might include: flight divers ion, declaration of emergency situation, eventually passenger evacuation, compartment inspection, fire ex tinguisher replacement, passenger disappointment, loss of confidence in the warning system etc ( Schmoetzer, 2001) Technologies Because the capability of available technology is seen as a significant constraining factor on the ability to screen, inspect, and track cargo, initiatives to improve cargo screening technology have been a focus of recent legislation to enhance air cargo security. Various techn ologies are under consideration for enhancing the security of air cargo operations such as: Tamper evident and tamper resistant packaging and container seals Cargo screening technology using x rays, chemical trace detection systems, or possibly neutron beams Canine teams Hardened cargo container technology Biometric technologies In addition, technologies to better track cargo shipments are being considered to maintain better control and tracking of cargo shipments along the supply chain. Both global positioning system (GPS) and radio frequency identification (RFID) technologies are seen as emerging technologies for improving the tracking of air cargo in the supply chain (Elias, 2007). And within that supply chain, there is a definite weak point during air transit. Therefore, there is a growing interest to know which cargo is on board, where is

PAGE 49

49 it in the cargo hold and what is its temperature, humidity, acceleration, etc (Schmoetzer, 2005). RFID in Aviation Cerino and Walsh (2000) stated that RFID tech nologies and systems with potential application to the worldwide aviation industry will most likely operate in the international Industrial, Scientific and Medical (ISM) frequency spectrum. Operation in an ISM band has the distinct advantage of not requiri ng the user to obtain specific licenses on a site by regulations on a band by band and geographical region by geographical region is required to allow completely unrestricted site operation anywhere within that ISM region In addition for the most part, the application of RFID for aviation has focused on as a replacement for a barcode baggage tag. However, the FAA and others have also addressed issues s uch as RFID for use with baggage containers, passenger and cargo tracking and as such have considered other additional frequencies; one such is the 915 MHz ISM band (Cerino and Walsh, 2000). The Interna tional Air Transport Association (IATA) member airlines unanimously approved the IATA Recommended Practice (RP) 1740C document, which endorses the use of ultra high frequency tags and readers compliant with the EPCglobal Gen2 protocol as a global air interface standard for RFID ISM F requency and A viation RFID C onsiderations The US Federal Aviation Administration (FAA) t o a large extent, and other aviation industry vendors and / or air carriers to a lesser extent, has completed testing of various technologies to ascertain their potential performance for aviation operational utility. Each system (being comprised of reader and tag) was not tested side by side, nor

PAGE 50

50 under identical simulated or operational conditions, yet a rather extensive matrix of frequency ( ISM frequencies of 125kHz / 132kHz, 13.56MHz and 2.45GHz ) versus operational performance was obtained as result of th e totality of FAA trials. In total, seven test phases geared at addressing the full range of aviation functional requirements alo ng with other additional site specific operational RFID usage evaluations contributed to the following results. Cerino and Walsh (2000) analyzed the results of three years of testing and research to develop an effective RFID system for the airline/airport environment. It was found that the 2.45GHz system has better performance, is more flexible in design, can be assembled off the shelf, and the system and tags (about 1/3 the cost of 13.56MHz disposable tags) are least expensive. The 13.56MHz system requires a mostly customized design, is less mature than the 2.45GHz technology, and interference concerns at 13.56MHz add signific ant complexity and cost to the system. The system at 125kHz / 132kHz presents significant tag cost disadvantages, which are not likely to be overcome for aviation use It is important to recognize that for each ISM band there exists specific transmitter po wer, signal modulation, duty cycle and other technical parameters that effectively band to band and (coupled with the natural differences in propagation characteristics for each band) result in significant differences as they relate to efficient aviation industry use. 2.45 GHz : At t his frequency the communication is entirely propagation coupling. Propagation in this band is via directional antennas, and hence reader energy ca n be

PAGE 51

51 directed to the area of greatest tag likelihood. The 2.45GHz band essentially evolved with the understanding that ISM interference exists at every band, and hence offers several advanced communications protocols which counter the potential interferenc e effects. Aviation Applications Terrorism, in particular its application to RF baggage tags and positive passenger bag matching, the use of RFID for commercial aviation extend beyond one security application. The fact is, there are many business and security reasons for applying RFID to the airport environment. Communications and RFID technology programs represent the key to integrating all th e components of airport secu rity, including perimeter intrusion detection, personnel screening, checkpoint screening, vehicle and cargo screening, digital video surveillance and recording, and RF ID baggage and vehicle tracking (Hallowell and Jankowski 200 5) Passenger baggage sortation In aviation industry, major ai rports have been looking for opportunities to use RFID technologies in baggage handling area s since 1999. Many pilot tests have been done at numerous airports including Gimpo (Korea), Las Vegas, Jacksonville, Seattle, Los Angeles, San Francisco, Boston, New York, Heathrow and Rome ( Chang et al, 2006 ) Ouyang et al. (2008) presented a n i ntelligent RFID r eader and its a pplication in the a irport b aggage h andling s ystem Jacinto et al. (2009) present ed an RFID equipment tuning and configuration methodology developed in a project to support baggage tracking and feed dashboards with real time status of Service Level Agreements between the airport, the airliner and the ground operators.

PAGE 52

52 A survey of the aviation industry would quickly identify that most air carriers utilize each individual baggage item. The aviation industry is working towards standardization of an RF baggage tag, in hope s of eliminating current barcode read rate limitations. These limitations become particularly evident with transfer baggage w h ere the use of automated sortation systems is essential to keep connection times to a minimum. T he FAA has, with the cooperation o f many vendors, airlines and airports, clearly demonstrated the ability to utilize RF baggage tags to enhance baggage sortation (Cerino and Walsh, 2000; Chang et al., 2006) Passenger / baggage matching. This process necessitates insuring that only boarde accomplished either totally manually or semi automatic (using handheld barcode). Either way, a significant amount of baggage handling is required and as such there exists a depend ence on human accuracy and barcode quality. The later becomes condition as when originally issued. In addition, certain low cost, high performance RFID tags afford th plate data. Information such as: the results of security screening; passenger biometrics; baggage images; and, flow timing through the baggage handling process all provide for significant co re business and security benefits (Cerino and Walsh, 2000). Verification / a uthentication From a security standpoint, few processes are considered complete without featuring a verification / authentication element. For example, if a certain passenger check

PAGE 53

53 used as basis for loading on to an aircraft. Consequently, it is imperative that the overall process be able to accurately verify that the bag in question definitely is the bag t hat has been deemed to be cleared for loading versus any other bag in the system. A robust RFID system implementation could easily track that bag based on its physical characteristics (with the support of other sensors); unique data (securely) added to t he tag and/or overall RFID carrier IT system; and, know path / timing for the bag through the baggage handling process (Cerino and Walsh, 2000). McCoy et al. ( 2005 ) investigated an automatic tacking system to improve airport security and efficiency by mean s of a cellular network of passive RFID receivers, combined with far field active RFID tags which may be issued within boarding cards or as security badges. Tracking and locating The tracking and rapid locating of baggage, cargo and containers (and the ass ociated integrity assurance of those items) is essential to the overall security of a the most part, only by a very labor intensive manual process. The application of RFID not only would be more time and labor efficient, but also more accurate. Sensors and electronic systems do not fall asleep, forget their assignments, become distracted, or otherwise perform in an unpredictable manner. The introduction of RFID to pr ovide this asset tracking and locating offers the opportunity for: centralized monitoring; continuous surveying; automatic event logging; and, of course, more rapid finding of items when retrieval is mandatory (Cerino and Walsh, 2000). Boeing and Airbus ar e also promoting the adoption of industry solutions for RFID on commercial aircraft parts. They believe that RFID could provide major benefits for the entire industry. They will get more accurate information about their demand for parts

PAGE 54

54 and will be able to reduce their parts inventory and cut the time it takes to repair planes. Part suppliers will also be able to reduce inventory, improve the efficiency of their manufacturing operations and reducing the amount of unapproved parts that enter the supply chain (Chang et al., 2006). C argo A natural evolution, beyond the sortation of passenger baggage, is the use of RF tags to handle sortation of cargo. For this application, the process is quite similar to passenger baggage sortation, with exception that cargo p arcels often can have a rather large form factor. As such, only an RFID system with enough flexibility and performance features to allow achievement of cargo and passenger baggage sortation requirements would be a realistic option for most of the aviation community. Naturally, for carriers such as FedEx and UPS this restriction do not directly apply, however, even in those examples, there is benefit in commonality with passenger air carrier systems. With tized shipments becomes a consideration. This could demand fixed station RFID readers requiring an even greater read volume or, more practically, implementation of an area read capability (such as the entire cargo hanger or even the entire tarmac area) to cover widely dispersed assets. This concept of broad coverage area for very large parcels also relates to passenger baggage, as in containerized luggage for stowage on wide body aircraft. And, as it relates to efforts accomplished and planned by the FAA, t he application of RFID to ensure the integrity (from a security standpoint) of ULDs from the time they are filled until the time they are loaded onto the aircraft (Cerino and Walsh, 2000). Containers. According to Cerino and Walsh (2000), passenger baggage containers (ULDs), are an important aspect of commercial aviation. For wide body

PAGE 55

55 aircraft, they are used to store up to 70 individual pieces of luggage (depending, of course, on the size of the individual pieces) and are loaded as a single unit. Whenever possible, baggage handlers segment passenger baggage by destination, in an effort to reduce the amount of handling time for luggage. This process of sorting luggage by destination and container positioning, besides supporting minimized ground turn around t ime, is important to support the security process of passenger bag matching. Accordingly, the use of RFID to: manifest exactly which baggage is in which container; locat e the container which holds the passenger baggage, in valuable from both security and operational efficiency standpoints. Monitoring. With air freight increasing rapidly; just in time delivery being a real challenge; ULDs being packed by third parties; ins pection time being very limited and transportation security being linked to volume, monitoring of goods within containers may contribute to security. Moreover, customers and insurance providers want to know what happens to their shipment (liability issue) and forwarders need to increase the monitoring of goods. Improvements may be feasible for temperature sensitive goods, hazardous materials and high value cargo, and such monitoring require communication between the container and the aircraft (Schmoetzer, 2 005). If readers are installed in the aircraft, they can be used to interrogate the cargo, acquire data from dedicated goods, enhance the monitoring, set off warnings, etc (Figure 2 7). Therefore a utomated onboard identification of cargo can contribute to enhance of security (Schmoetzer, 2005) Benefits of RFID assisted air freight handling include: Automated tracing of goods Automated verification of aircraft load instruction

PAGE 56

56 Reduction of false loading Reduction of ground time Paperless data transfer Elec tronic Bill of Loading Wireless interface can be used to enable more services than simply RF Identification e.g.: o Change / update information on relevant item o S elf control / monitoring means o M emorize what i s of interest o D ata exchange (e.g. actual temperatu re, history, etc.) Cold c hain Products, such as food, pharmaceuticals and flowers, are at high risk of perishing from various adversities along the cold chain. The parties involved should control when possible, and at the very least monitor the conditions of the goods in order to ensure their quality and to comply with all legal requirements. Among environmental parameters during transport, temperature is the most important in maintaining the shelf life of the products (Nunes et al., 2006; Zweig, 2006; Jed ermann et al., 2009). delivered year round all over the world, thus, temperature sensitive items are likely to be shipped by air because of their relatively short shelf life. Unf ortunately, a faster transit time does not always imply controlled temperature throughout transportation. In contrast, during airport operations, loading, unloading, air transportation or warehouse storage, perishable goods often suffer from temperature ab use either due to difficulties in controlling the temperature, absence of refrigerated facilities, or lack of information about produce characteristics and needs (Nunes et al., 2003). O f approximately 2.6 million tons of perishables air freighted in 2008, nearly 30% is estimated to be lost due to handling and temperature abuse (Catto Smith, 2006). In a previous study, mond et al. (1999) showed that the environmental conditions during airport operations could, in

PAGE 57

57 fact, be very far from the optimum for fruit s and vegetables. Moreover, in a strawberry quality study, Nunes et al. (2003) showed that greater losses in quality occurred during simulation of the airport handling operations, in flight, and retail display than during warehouse storage at the grower, t ruck transportation to or from the airport, or during backroom storage at the supermarket. The relative success of growing exotic perishables in countries such as Colombia, Ecuador, and Peru in South America, and its successful distribution and commerciali zation in distant markets such as the US and the EU, have been made possible due to advances in transportation and refrigeration technologies. Yet the transportation systems of exotic perishables are far from perfect. Transportation and logistics costs can be high both monetarily and in terms of loss of quality during handling (Vega, 2008). While passenger business is generally bidirectional, cargo is not. Consequently, freighter routes are often imbalanced. This implies that when transporting goods from po int A to point B, the freight rate charged must also cover the return trip from B to A (Vega, 2008). Temperature monitoring Currently, most digital temperature loggers have to be connected to a host device to download data, and as a result, have limited real time data interactivity, which result s in after the fact analysis for claims, loss in quality and related issues. Radio frequency identification (RFID) temperature loggers function wirelessly which allows for real time information transfer. Active or semi passive RFID tags can support one or many sensors as well as the unique ID that RFID technology provides by design. The RFID tag, with associated hardware and software add s the benefit of having the item scanned on receipt, so that if an alert (alert triggers are

PAGE 58

58 programmable prior to shipping) is active, the receiver knows immediately (not after the fact) that there is a potential problem with the shipment and can spend the time required on specific shipments rather than going through random inspectio ns (Jedermann et al., 2007). Many studies have already shown the effectiveness of RFID in monitoring product temperature during transit ( mond 2007; Jedermann and Lang, 2007; Jedermann et al., 2007; Ketzenberg and Bloemhof Ruwaard, 2009). Wireless Interfe rence concern during flights. For various reasons, many devices such as laptop computers are allowed during flights, while intentional transmitters such as wireless device s and phones are prohibited. There have been many past studies addressing the three elements. Examples include emissions measurements from wireless devices in aircraft radio bands (Nguyen et al., 2004, 2005). Aircraft radio receiver interference thresholds data may be found in "Environmental Conditions and Test Procedures for Airborne Equipment" (DO 160 F ) published by the Radio Technical Commission for Aeronautics (RTCA) Devereux et al. (1997 b ) reported research and experimental results of commercial aircr aft avionics and control systems exposed to conducted and radiated el ectromagnetic interference This experiment, conducted inside a CV 580 aircraft, determined how susceptible installed avionics are to different low power RF sources located in passenger c abin and baggage compartments, and avionics and cargo bay areas. This study show ed that avionics certified to the special 100 V/m high intensity radiated field requirements are still highly vulnerable to low to moderate levels of onboard RF energy when inst alled in the aircraft system. V arious sources of

PAGE 59

59 electromagnetic interference can be easily transmitted at these frequencies. Also, many of the avionics bays and cable routes onboard today's modern aircraft are unshielded and easily accessible. Many of the aircraft cargo bays are located adjacent to the avionics bays with only fiberglass walls and access doors separating the two ( Cerino et al., 1997 ) A cargo hold could easily contain a RF transmitter system which emits enough RF energy to seriously interfe re with and upset today's avionics systems. Further analysis of the CV 580 aircraft was made by Devereux et al. (1997 a ) which emphasizes the aircraft receiver in band frequency susceptibility, primarily the microwave frequency bands used for aircraft Globa l Positioning Satellites (GPS) and Satellite Communication (SATCOM) navigation. This report also includes VOR, VHF, UHF and DME navigation and communication frequency bands (refer to Table 2 3) The evaluation is focused on coupling behavior from inside th e aircraft cavity, coupling through the windows, to the aircraft receive antennas It was found that path loss coming from the cargo bay is likely to travel up through the mostly non metallic ceiling of the cargo bay into the passenger cabin and out the pa ssenger windows or through the non metallic avionics bay doors located in the cargo bay through to the cockpit window and to the receive r antennas. Moreover, the path losses from a transmitting antenna at a central position inside the fuselage were greater than that from any window and about 10 to 20 dB greater than from the optimum window. Furthermore, in a study by Nguyen et al. (2007) i nterference coupling factor (or interference path loss) data were measured for multiple radio systems on ten small airc raft. The data show significant data variations between different aircraft models. The

PAGE 60

60 data also show stronger interference coupling than for previously measured larger aircraft, and potentially result in higher interference risks. Electronic devices Unli ke aircraft installed equipment, passenger carry on devices such as wireless phones are not required to pass the rigorous aircraft radiated field emission limits. Previous studies were made to measure the emissions from wireless phones in aviation bands an d to assess interference risks to aircraft systems (Ely et al., 2003; Nguyen et al., 2004). Results from a recent study showed that the spurious emission s from 33 phones tested were below the aircraft installed equipment limits (RTCA/DO 160 cat. M), even w ith the consideration of the 5 8dB uncertainty associated with the phones expected directivity (Nguyen et al., 200 5 ). RFID interference Passive tags are considered less of an interference concern for aircraft since they do not transmit without an interroga tor, whose electromagnetic fields power the tags. A ctive tags, on the other hand, are powered with internal batteries. Active tags can be of higher interference risks since many can transmit on their own without an interrogator. The actual interference ris ks depend on several and unintentional emission levels, the propagation path loss factor, and the victim type. Nguyen et al. (2008) studied the emission measurement s of active tags and their interference potential on aircraft sensitive radio receivers. Specifically, this study measure d the unintentional emissions from several popular RFID tags used for cargo tracking. Results showed that many d power exceeded RTCA/DO 160E category L and M EIRP emission limits, one of which surpassed the limit by as much as 35 dB in the GS band

PAGE 61

61 (328.60 335.40MHz). Another study made with two active tags at 433 MHz showed that e mission s from both tags were highe r than the limit of DO 160E at the operation frequency and the harmonics. It was also found that the emission level depends greatly on the location and material on which a tag is placed (Yonemoto et al., 2007) For a complete interference assessment, other factors such interference path loss and receiver interference thresholds should also be considered. These factors were addressed previously for example, Nguyen et al. (2006) documented the measurement of interference path loss for cargo bays on a Boeing 747 and an Airbus A320 aircraft. Nguyen et al. (200 4 ) provided a summary of passenger cabin path loss data for many commercial transport aircraft. Nguyen et al. (2007) reported the path loss measurements for general aviation aircraft. These path loss data represent the propagation loss between the tag Aircraft radio receiver interference thresholds for continuous interference signal transmission were addressed in RTCA/DO 294B (Guidance on allowing transmitti ng portable electronic devices (T PEDs) on aircraft) LaBerge (2007) analytically determine d thresholds for intermittent interference signals similar to RFID emissions. Moreover, t he work in Nguyen and Mielnik (2008) reports the laboratory effort to determ ine the GS system interference threshold to an RFID interference signal. RFID airworthiness policy The current Federal Aviation Administration (FAA) RFID policy document is AC20 2008). It incorporates and supersedes jointly issued AIR 100, AFS 200 and AFS 300 RFID policy. This advisory circular offers guidance on inst alling and using RFID systems on aviation products and equipment. Specifically, it provide s an acceptable way to us e

PAGE 62

62 RFID readers or interrogators installed on aircraft, and advice on allowing use of RFID devices on baggage, mail containers, cargo devices and galley / service carts. It also covers using portable RFID readers or interrogators carried onboard aircraft. T his advisory circular is not mandatory and does not constitute a regulation. It describes an acceptable means, though it i s not the only means, to show compliance with applicable installation and operational requirements. Through this AC, the FAA does not prohibit any use of RFID devices, it provides specific requirements that the equipment must meet to be safe and airworthy. Concluding R emarks This review showed that RFID has long been thought to have the potential to help the air cargo industry by increa sing its safety, operation efficiency, monitoring capability, customer satisfaction, etc. Before implementation can take place, some questions have to be answered and more research has to be done This is why this study will cover the following subjects: i nterference and frequency assessment in air cargo warehouses ; RF propagation study inside the cargo hold; RF propagation through ULD materials; and temperature distribution in the cargo hold during transit.

PAGE 63

63 Table 2 1 EPC global tag class structure EPC Cl ass Functionality Type Class 0 Gen 1 Read only Passive Class 1 Gen 1 Write once, read many Passive Class 1 Gen 2 Write many, read many Passive Class 2 Write once, read many Passive Class 3 Read and write Semi passive Class 4 Read and write Ac tive Class 5 Reader tags Active Table 2 2. Permissible field strengths for RFID systems in accordance with FCC Part 15 (FCC, 2008). Frequency range (MHz) Max. E field (mV/m) Measuring distance (m) 13.553 13.567 10 30 433.5 434.5 902.0 928.0 11 50 3 3 2435 2465 50 3 5785 5815 50 3 Table 2 3. Aircraft radio systems Aircraft Band Description Receive Spectrum (MHz) LOC Localizer 108.10 111.95 VOR Very high frequency omnidirectional range 108.00 117.95 VHF Com Very high frequency voice communication 118.00 138.00 GS Glideslope 328.60 335.40 DME Distance measuring equipment 962.00 1213.00 ATC Air Traffic control radar beacon system 1030.00 TCAS/TCAD Traffic collision avoidance system / Traffic collision alert device 1090.00 GPS (L5) Global positioning system 1176.45 GPS (L2) Global positioning system 1227.60 GPS (L1) Global positioning system 1575.00 2 MLS Microwave landing system 5031.00 5090.70

PAGE 64

64 Figure 2 1. Electromagnetic spectrum

PAGE 65

65 Figure 2 2 Different par ts of a wave (Lahiri, 2006) Figure 2 3. Example of an RFID system on conveyor belt. Figure 2 4. Wave propagation for linear and circular polarization

PAGE 66

66 Figure 2 5. Cargo warehouse floor plan and activity areas.

PAGE 67

67 Figure 2 6. International fligh t temperature profile with both high temperature excursions (during stopovers) and low temperature excursions (in flight) ( Syversen et al., 2008) Figure 2 7. Example of a GSM/GPS capable RFID system for real time data acquisition (Schmoetzer, 2005).

PAGE 68

68 CHAPTER 3 AIR C ARGO WAREHOUSE ENVIR ONMENT AND RF INTERFERENCE Introduction The air cargo system is a complex network that handles a vast a mount of freight, packages, and mail carried aboard passenger and all cargo aircraft. With air freight increasing rapi dly; just in time delivery being a real challenge; unit load devices ( ULDs ) being packed by third parties; inspection time being very limited and transportation security being linked to volume, monitoring of goods within containers may contribute to enhanc ed security (Schmoetzer, 2005). Moreover, monitoring the movement of goods within the air cargo warehouse can enhance operation efficiency and provide valuable real time information. Cargo warehouse A typical cargo terminal is divided into an import area and an export area. The import area is dedicated to receiving, processing and releasing inbound freights. The export area is dedicated to receiving, processing and preparing outbound freights. The flow of goods through the terminal is either from the airsi de to the landside (terminating freights or connecting freights requiring the road feed service), from the landside to the airside (originating freights or connecting freights arriving from a road feeder service), or from the airside to the airside via the terminal (connecting freights). These movements of goods in many directions can easily lead to item misplacement, and therefore it can be time consuming for the air cargo agents to locate the pieces when they are needed for build up. Item level radio freq uency i dentification ( RFID ) tagging, as well as ULD tagging, can address this issue by allowing real time access to location of all shipments present in the warehouse. Moreover, RFID could permit automated and efficient ULD / item association during contai ner build up. This

PAGE 69

69 technology could also give time stamp data on the movement of ULDs in and out of the building. There are many applications for which RFID could be helpful in an air cargo warehouse, but for a reliable system to be working properly, inter ference assessment needs to be done. RF i nterference Interference is typically interpreted as other sources of radio frequency (R F ) energy that compete with an existing implementation Any devices operating at the same frequency have the potential to inte rfere with each other. There exist many short range consumer devices operating at UHF frequencies. For example, devices like automotive remotes alarm systems, home automation and wireless temperature sensors can operate at 433MHz. At 915MHz, GSM 900 is a potential threat, as well as cordless phones, stereo, older wir eless local area network (WLAN) amateur radio etc. At 2.45GHz, interference is possible from devices such as new er WLANs, cordless phones, microwave ovens, fluorescent lighting and Bluetooth technology to name a few. There are ways to work around this issue by setting up these systems so that they do not use the sa me RF channels at the same time, but their consideration as potential interference threat to any UHF RFID system is of prime impor tance. Frequency allocation RFID is a radio technology and as such requires the use of radio spectrum to operate. Generally, when discussing spectrum issues, focus tends to be on tags that make use of the ultra high frequency ( UHF ) range UHF (together w ith VHF) is the most common frequency band for television. In addition, it is used for mobile telephony, two way radio communication and increasingly for digital services. Since RFID shares the UHF range with other applications, only a limited bandwidth wi thin

PAGE 70

70 UHF is available for s hort range devices and RFID, which is 902 928 MHz in North America Although there are RFID tags and applications that make use of other frequency bands ( for example 433MHz and 2.45GHz ), most spectrum issues seem to concentrate a round the UHF range and many sectors and applications (e.g. in retail and supply chains) use UHF tags (van de Voort and L igtvoet, 2006). ISM band The In dustrial, S cientific and M edical (ISM) band s are defined by the International Telecommunication Union Recommendation ( ITU R ) in 5.138, 5.150, and 5.280 of the Radio Regulations. Individual countries' use of the bands designated in these sections may differ due to variations in national radio regulations. Because communication devices using the ISM bands mu st tolerate any interference from ISM equipment, these bands are typically given over to uses intended for unlicensed operation, since unlicensed operation typically needs to be tolerant of interference from other devices anyway. In the US the use of ISM bands is governed by Part 18 of the FCC rules, while Part 15 Subpart B contains the rules for unlicensed communication devices, even tho se that use the ISM frequencies All frequencies considered in this study are part of the following ISM bands: 433.05 43 4.79 MHz 902 928MHz (North America only) and 2400 2500MHz Spectrum analyzer The sensitivity (or threshold) of a spectrum analyzer is defined as its ability to detect signals of low amplitude. The maximum sensitivity of the spectrum analyzer is limited by the noise generated internally (Anritsu, 2008 ). Noise level is directly proportional to the resolution bandwidth (RBW) of the system Therefore, by decreasing the bandwidth by an order of 10dB m on the logarithmic scale (from 100KHz to 10KHz, for instance ) the system noise floor is also decreased by 10dB m. As

PAGE 71

71 an additional example, when the RBW is increased from 100Hz to 10kHz, the noise floor moves up 20dBm ( Figure 3 3 ). When comparing spectrum analyzer specifications it is important that sensitivity is c ompared for equal bandwidths since noise varies with bandwidth. For this test, all data were taken at 10kHz RBW, which was the maximum sensitivity achievable with this specific spectrum analyzer. In other words, the real noise floor could have been lower t han what was recorded. Reader sensitivity RFID communication go es two ways, first the reader tag link, then the tag reader link. The reader sends an RF signal with initial output power, which attenuates while traveling in the medium between the reader and the tag. If the tag received enough signal to respond, it sends its info back to the reader, which is also attenuated on the way back. When the reader receives this much attenuated signal, if it is above reader sensitivity levels, the communication is com plete. The reader sensitivity defines the minimum signal level needed to be able to communicate with the transmitter. For passive R FID, it is usually the reader tag link that is the limiting one, whereas f or a semi passive tag, the reader tag link requirem ent is much more lenient since the received power must only be decoded not exploited (Dobkin, 2008). Reader sensitivity is dependent on several design choices, particularly, antenna configuration and will become more important as tag IC power is scaled to lower values. Moreover, when interfere nce is present, it is the tag reader link or reader sensitivity that limits the system the most. The sensitivity of the receiver is limited by the noise that enters it and the largest source of noise for an RFID recei ve r is usually the leakage from its own transmitted signal (Dobkin, 2008). Nevertheless, any other sources of interference, within the specific UHF band used, can affect the performance of an RFID system when

PAGE 72

72 its noise level is above the reader sensitivity limit. For example, a few years ago the typical passive RFID reader had a sensitivity of 65dBm, whereas in 2007 the most sensitive reader had a sensitivity of XC3 reader has a receive s ensitivity of 120 dBm (Intelleflex, 2010). This does not mean that any noise above 120dBm would completely disable the communication link between a tag and reader. Interference can only be expected in the case where the noise level would be well above the tag reader link; and that is not taking into account the fact that most RFID readers have their own interference filtering system. In other words, each RFID reader will tolerate up to a certain level of signal to noise ratio (SNR in dB) which indicates how much higher the signal level is compared to the noise level. For instance, if a reader with sensitivity Y dB requires an SNR of X dB for a successful read, this means that in order to use the reader at its maximum capacity, the noise level present in the environment has to be less than (Y X) dB. If this is not the case, than the effective distance between the tags and the readers need to be adjusted accordingly to compensate for the environmental noise. For RFID technology to be implemented into an air cargo warehouse, many RF assessments need to be done. In particular, interference levels might dictate which frequency is most suitable for a specific RFID application. T herefore, t he objective of this study was to identify the multiple RF int erferences encountered inside air cargo warehouse s Material s and Methods RF interference readings were taken at different locations throughout two air cargo warehouses. One warehouse was located near the Pierre Elliott Trudeau airport (YUL) in Montreal, Canada, whereas the other one was near Pearson airport (YYZ) in Toronto,

PAGE 73

73 Canada. All interference readings were taken with a handheld spectrum analyzer (HSA 9101, Willtek, Parsippany, NJ ) at three different UHF frequencies: 433MHz, 915MHz and 2.45GHz. Each frequency was measured using a specific receiving antenna, as detailed in Table 3 1. Signal data w ere adjusted to account for receiving antenna gain. In other words, readings at 915MHz were lowered by 2.5dBm and readings at 2.45GHz were lowered by 8dBm. All interference data are in dB m and the range considered for each frequency is in accordance with the F ederal C ommunication C ommission part 15 regulations (FCC, 2008). For 433MHz, the range of device operation is between 433.5 and 434.5MHz; for 915MHz, the range is between 902 and 928M Hz; and for 2.45GHz, the range is from 2.435 to 2.465GHz. Export : All freight arrives at the cargo facility either as bulk or as shipper loaded unit device s (SLUD), which can be air container s or pallet s pre loaded by the customer. The freight gets weigh ed and dimensioned by the acceptance agent and stored at the appropriate location depending on its flying time and destinatio n. If items are bulk, they ultimately go to the build up area to be put in a ULD (pallet or container) or are transported in a tub cart directly to the airplane if this airplane is bulk loaded. ULDs are transported onto roller systems or forklifts through the cargo facility, after which they are carried out to the ramp onto trailers. Import : ULDs are transported between the aircraft and cargo facility on special trailers pulled in a train by small tractor vehicles. Bulk items are unloaded from the aircraft directly into tub carts. Everything is brought back to the cargo warehouse and is broken down when needed and stored until custome r pick up.

PAGE 74

74 Montreal W arehouse T his location is in a very new building (constructed in 2008) and covers an area of 16,300m 2 Figure 3 1 presents the layout of the warehouse, including most activity areas (export, import, storage, build up, break down, airsi de and landside). Interference reading points are circled on the layout and correspond to the following areas: Office area Export landside, near unloading dock doors Breezeway, export airside, near doors Breezeway, import airside, near doors Center of ware h ouse, between import and export; pallet break down and build up area ; and storage Import landside, near loading dock doors Toronto W arehouse T his location is slightly older building, built in 2002, and covers an area of 26,700m 2 which is more than 60% b igger than the Montreal warehouse. Figure 3 2 presents the layout of the warehouse, including most activity areas (export, import, storage, build up, break down, airside and landside). Interference reading points are circled on the layout and correspond to the following areas: Import landside, near loading dock doors Office area Export landside, near unloading dock doors Build up area Center of wareh ouse, between import and export Break down of imported ULDs Breezeway, import airside Middle of breezeway Bre ezeway, export airsi de Results and Discussion As stated earlier, frequency ranges are as follow s : 433.5 to 434.5MHz; 902 to 928MHz; and 2.435 to 2.465GHz. Table 3 2 and 3 3 show minimum and maximum

PAGE 75

75 electromagnetic signal levels recorded inside these range s for Montreal and Toronto area (Table 3 2 and 3 base for peak signal identification. I t represents the sensi tivity of the spectrum analyzer. When minimum and maximum data points are similar for a single location, it can be above noise floor. 433MHz Around 433MHz, the study d id not show any important signal peaks that could interfere with an RFID system operating in this range. The average minimum background noise is 107.3dBm and the maximum sign al level recorded is 10 3. 7 dBm in Montreal for position 1 and 3 (Table 3 2); and 8 9 8 d Bm in Toronto, position 4 (Table 3 3). Figure 3 4 shows the spectrum analyzer graphs for 433MHz in Toronto, position 4. Only one graph per frequency has been chosen as illustration, which represents the worst case scenario recorded within all locatio ns and positions at that frequency. A ccording to the National Telecommunications & Information Administration (NTIA, 2003, 2010) the small signal peaks visible around 433MHz (Figure 3 4) corresponds to Amateur (Capital letters are prima ry activity and lower cases are secondary) in the range of 420 450MHz. The ISM range 433.05 434. 79MHz can be heavily occupied by a wide range of ISM applications. In addition to backscatter (RFID) systems, baby intercoms, telemetry transmitters (including those for domestic applications, e.g. wireless external thermometers), cordless headphones, unregistered LPD walkie talkies for short range radio, keyless entry systems (handheld transmitters for vehicle central locking) and

PAGE 76

76 many other applications are cra mmed into this frequency range. Unfortunately, mutual interference between the wide range of ISM applications is not uncommon at this frequency ( Finkenzeller, 2003). However, as far as the results found in this experiment are concerned, the very minimal pe ak recorded as the worst case scenario does not cover the entire band and is not present at a high power level. It is in fact lower than reader sensitivity of RFID readers designed three years ago. Therefore, it would be safe to say that the level of inter ference recorded would not significantly affect the performance of an RFID system. 915MHz As shown in Figure 3 5, there is very low activity within the 902 928MHz range in the Toronto warehouse, which represents again the worst case scenario recorded. The maximum signal level recorded was 99.8dBm for Montreal and 92.2dBm for Toronto, whereas the average minimum is 109.5dBm (Table 3 2 and 3 3). As explained earlier, this level of noise could possibly only interfere with state of the art readers that have sensitivities below 90dBm and are being used at their maximum capacity (maximum range which allows for successful communication). In the event that this level of noise would not be blocked or filtered by the reader, it would still only interfere with tags that are further away from the reader, and require this much sensitivity to communicate. Typically, when the interference level is close to the reader sensitivity, it can only affect erence noise and the receiver sensitivity. Moreover, systems working in that frequency range in North America use frequency hopping techniques; which means that the reader utilizes a slim portion of the band for a maximum of 0.4 seconds at a time. In other words, if the

PAGE 77

77 interfering signal is only present on a small portion of the band, the likelihood of interference becomes lesser. As seen in Figure 3 6, activity surrounding this band corresponds to (NTIA, 2003, 2010), where capital letters are primary acti vity: 85 1 89 4: FIXED + LAND MOBILE (GSM 850) 92 9 93 2 : FIXED + LAND MOBILE 932 935: FIXED 935 940: FIXED + LAND MOBILE Wireless applications that operate license free in the 900 MHz ISM band include superviso ry control and data acquisition industrial aut omation, building automation and control, wireless sensor networks and consumer devices such as cordles s telephones, wireless speakers and baby monitors. In the U.S., most such systems utilize frequency hoppi ng spread spectrum over 902 928 MHz. 2.45GHz Ar ound 2.45GHz, the study shows much more activity that could interfere with an RFID system operating in this range (Figure 3 6). The max imum signal level recorded is 74 6 dBm in Montreal fo r position 4; and 66 .5dBm in Toronto for position 9 w hereas the av erage minimum is 112.6dBm (Table 3 2 and 3 3). RFID readers working at this frequency have similar sensitiviti es as readers for 915MHz and frequency hop within the specified range. As mentioned earlier, frequency hopping only allows a small portion of th e band to be used at a time; therefore if the interfering signal is only present on a portion of the band, the likelihood of interference becomes smaller. Signal level of around 66 dBm is well above most reader sensitivity, which could cause unwanted inter ference in the event that those readers are not capable of filtering the interfering signal.

PAGE 78

78 The signal peaks visible around 2.45GHz for Montreal (Figure 3 6) corresponds to (NTIA, 2003, 2010), where capital letters are primary activity and lower cases ar e secondary: 2417 24 46 : Amateur + radiolocation 245 4 24 72 : FIXED + MOBILE + radiolocation In the 2.4GHz ISM band there are several sources of interfering signals, including but not limited to: microwave ovens, baby monitors, wireless phones, wireless came ra, Bluet ooth enabled devices, W LANs WIFI, and 2 way radios. Conclusion This study showed interference levels at three UHF frequencies recorded in two air cargo warehouses. The interference levels from highest to lowest were at 2.45GHz, 433MHz, and 915MHz respectively. Even in the case where interference is above a typical reader sensitivity, it is hard to say that these signals will or will not interfere with and RFID system since every system is designed differently. When the interference is on the edge of the spectrum, it is possible to reduce the used RF band in order to avoid such noise. When the interference is more towards the middle or across the entire band, RFID system design must take this into consideration to filter the noise out. RFID systems working at 915MHz and 2.45GHz use frequency hopping, therefore, when noise is only present on a certain part of the band, interference is also only encountered interference for a short period of time (depending on the percentage of the band covered by the interfering signal). This test was performed inside two warehouses, which may or may not give a realistic average representation of any air cargo warehouse in North America. But as far as these locations are concerned, it would be

PAGE 79

79 safe to say that implementation of RFID systems at 915MHz or 433MHz would bring the best results, interference wise.

PAGE 80

80 Table 3 1. Receiving antenna specifications Frequency Polarization Gain Model & Manufacturer 433 MHz Linear (omni) 0 dBi B 368 1, How Tsen Intl. Electronics Metal Co.,Ltd. Shin Wu Hsiang, Tao Yuan Hsien, Taiwan 915 MHz Linear (omni) 2.5 dBi EXR902TN, Laird Technologies, Schaumburg, IL 2.45 GHz Linear (omni) 8 dBi MRN 24008SM3, AntennaWo rld, Miami, FL

PAGE 81

81 Table 3 2. Minimum and maximum interference re adings (in dBm) for six positions and three frequencies at the Montreal warehouse. Frequency Position in the warehouse (MHz) 1 2 3 4 5 6 min max min max min max min max min max min max 4 33 107.3 103.7 107.8 103.8 106.6 103.7 106.7 104.4 108 104.5 108.1 105 915 109.8 103.8 109.8 100.0 109.0 101.9 109.0 99.8 108.6 100.4 108.6 105.9 2450 113.7 82.9 113.9 79.3 113.1 76.1 112.6 74.6 112.4 82.0 113.7 88.7 Table 3 3 Minimum and maximum interference readings (in dBm) for nine positions and three frequencies at the Toronto warehouse. Frequency Position in the warehouse (MHz) 1 2 3 4 5 min max min max min max min max min max 433 106.6 103.9 106.4 104 .2 107.1 103.8 107.4 89.8 106.7 104.5 915 110.5 99.9 110.1 103.7 108.9 95.9 110.1 98.1 110.5 94.4 2450 112.0 90.0 111.8 91.3 112.1 86.5 112.4 76.1 112.0 80.5 Table 3 3 Continued Frequency Position in the warehouse (MHz) 6 7 8 9 min max min max min max min max 433 107.8 104.2 107.2 105.3 107.3 101.8 108.2 103.9 915 109 98.4 110.1 102.2 110.1 92.2 108.9 93.4 2450 112.1 90.1 112.4 69.4 112.5 71.6 112.5 66.5

PAGE 82

82 Figure 3 1. Montreal cargo warehou se facility floor plan and interference reading points (numbered 1 to 6 ).

PAGE 83

83 Figure 3 2. Toronto cargo warehouse facility floor plan and interference reading points (numbered 1 to 9 ). Figure 3 3. Noise floor of spectrum analyzer at three different resol ution bandwidths.

PAGE 84

84 Figure 3 4. Worst case scenario for s ignal interferen ce readings around 433 MHz ( Toronto position 4 ) Span: 1 0MHz RBW: 10kHz a ttenuation: 0dB gain : 0dBi Figure 3 5. Worst case scenario for signal interference readings around 915 MHz ( Toronto position 8). Span: 50MHz, RBW: 10kHz, a ttenuation: 0dB gain: 2.5dB i

PAGE 85

85 Figure 3 6. Worst case scenario for signal interference readings around 2450MHz ( Toronto position 9). Span: 50MHz, RBW: 10kHz, a ttenuation: 0dB gain: 8dB i

PAGE 86

86 CHAP T ER 4 R ADIO F REQUENCY PROPAGATION INSIDE T HE CARGO HOLD OF A D C 10 AIRCRAFT Introduction New t echnologies to better track cargo shipments are responsible for maintain ing improved control and tracking along the supply chain. Both global positioning system ( GPS) and radio frequency identification (RFID) technologies are emerging technologies for improving the tracking of air cargo in the supply chain (Elias, 2007) Within that supply chain, there is a definite weak point during air transit. Therefore, there i s a growing interest to know which cargo is on board, where it is in the cargo hold and what some variables such as its temperature, humidity, acceleration, etc, are. (Schmoetzer, 2005). With air freight increasing rapidly; just in time delivery being a re al challenge; customers and insurance providers want to know what happens to their shipment (liability issue) and forwarders need to increase the monitoring of goods. Improvements may be feasible for temperature sensitive goods, hazardous materials and hig h value cargo, and such monitoring require s communication between the container and the aircraft (Schmoetzer, 2005). If RFID readers are installed in the aircraft, they can interrogate the cargo, acquire data from dedicated goods, enhance the monitoring, s et off warnings, etc. Therefore a utomated onboard identification of cargo can contribute to enhance d security (Schmoetzer, 2005) RFID airworthiness policy The current Federal Aviation Administration (FAA) RFID policy document is AC20 approval and operational allowance issued AIR 100, AFS 200 and AFS 300 RFID policy. This advisory circular offers guidance on installing and using RFID systems on aviation products and eq uipment. Specifically, it provide s

PAGE 87

87 an acceptable way to use RFID readers or interrogators installed on aircraft, and advice on allowing use of RFID devices on baggage, mail containers, cargo devices and galley or service carts. It also covers using portabl e RFID readers or interrogators carried onboard aircraft. The FAA airworthiness concerns about RFID system s installed on aircraft include: Integrity, accuracy, and authenticity of both safety related and identification data from RFID devices Fire and elect rical safety, crashworthi ness, and environmental effects RFID device generated RF intended transmissions or spurious emissions, both of which can interfere with aircraft electrical and elec tronic systems and components Maintenance requir ed for RFID devices and readers Therefore, some of the current requirements that passive and/or active RFID must meet are: Safety assessment Major alterations ( if it might structural strength, performance, flight characteri stics, or other qualities affecting airworthiness ) Electromagnetic compatibility (EMC) demonstration (for active RFID and readers) Battery safety (for active RFID) Flammability and fire safety Mounting and attachment integrity I nstructions for continued ai rworthiness (documentation) This advisory circular is not mandatory and does not constitute a regulation. It describes an acceptable (though not the only) means, to show compliance with installation and operational requirements. Through this document the FAA does not prohibit any use of RFID devices, instead it provides specific requirements that the equipment must meet to be safe and airworthy. RF Propagation While previous research (Rappaport and McGillem, 1989; Valenzuela et al., 1997; Mayer et al., 2006) documents indoor prop agation of radio waves, very little work (Laniel et al., 2009) specializes on RF behavior inside a metal

PAGE 88

88 environment. In contrast, the behavior of radio frequency around metal has been studied extensively (Dobkin and Weigand, 200 5; Griffin et al., 2006; Prothro et al., 2006; Sydanheimo et al., 2006). Because aluminum is a very good conductor, incident electromagnetic wave totally reflects from the metallic surface with a phase reversal (Cheng, 1993; Reitz et al., 1993). Such mater ials are generally referred to as being RF opaque Moreover, metallic surface of the object in the vicinity of an antenna changes its radiation pattern, input impedance, radiation efficiency and resonant frequency. These changes depend on the size and sh ape of the metallic object and also on the distance of the antenna from the object (Raumonen, 2003; Mo and Zhang, 2007). Mo and Zhang (2007) also demonstrated that RFID tags placed 1/4 wavelength away from the metallic surface enhances the readability of t he tags. On the other hand, little or no reflection occurs when electromagnetic waves penetrate directly through objects such as paper, non conductive plastics or textiles (Penttil et al., 2006). These materials, including most composites, are non absorbi ng and possess low refractive indexes. Such materials are generally referred to as being RF lucent Aircraft Whether it is inside the cargo hold or the cabin, RFID installed inside an aircraft would encounter a lot of metal in its environment. The entir e fuselage of most aircr afts is made of aluminum alloy Even if the use of composite materials is continuously growing and that big steps forward have been reached on the newest aircrafts, the material distribution on an aircraft structure predominantly re mains aluminum based alloys. For example, only a mere three to four percent of the original Airbus A300 was made of composites but they now account for 25% of the A380 structural weight and will account for more than 50% on the future A350 (Airbus, 2009).

PAGE 89

89 The shape of the metal enclosure is not only cylindrical, but also has many cross beams as shown in Figure 4 1 which creates a very uneven reflective surface. This metal environment leads to highly unpredictable RF propagation behavior. DC 10 30F Accord ing to Boeing (2010 c ), t he multi range DC 10 was designed and built in Long Beach, California, by Douglas Aircraft Company, now the Long Beach Division of Boeing Commercial Airplanes. Production was started in January 1968 and extended to 1989, where 386 c ommercial DC 10s were delivered. The DC 10 Series 30F, an all freighter model, was ordered by Federal Express in May 1984. T his pure freighter version carries palletized payloads of up t o 79,380kg on more than 6,115 km This is the model used for study in t his chapter (Figure 4 2 ). RFID is becoming more and more accepted for air cargo applications Moreover, RF propagation inside a ircrafts is not well documented. As a result, t he objective of this study is to evaluate the RF propagation behavior inside the c argo hold of a wide body aircraft (DC 10 30F) at different frequencies. Materials a nd Methods Three radio frequencies (433MHz, 915MHz and 2.45GHz) were tested inside the forward lower cargo hold of a DC 10 30 freighter aircraft (Figure 4 3). The aircraft cargo door was kept open during the entire testing period, which simulates loading or unloading environment. Each frequency was generated by either an RFID reader or an RF transmitter (Table 4 1). Each RF system was connected to its respective set of anten nas as described in Table 4 1. The RF systems were installed inside the cargo hold and the corresponding emitting antenna, connected via cable, was positioned either at the front of the cargo hold (top end), or in the center of the ceiling (Figure 4 4). On ly one frequency and one antenna position was tested at a time. The only thing

PAGE 90

90 present inside the cargo hold during testing was the RF testing equipment and one person, who was standing on the step just outside the cargo door while operating the spectrum a nalyzer. Details of the cargo hold and cargo door dimensions can be seen in appendix A. Test 1: Propagation Study Signal strength data was measured in dBm (power level in decibels relative to 1 mW) via a spectrum analyzer (RSA3303B, Tektronix, Beaverton, O R) connected to the appropriate RF receiving antenna (Table 4 1) and a 50m long LMR 400 low loss cable. The receiver antenna was mounted on a plastic tripod, which was moved every meter along the length of the cargo hold at three different height and three different width po sitions. This created a 3x3x 12 signal strength data grid for each frequency, antenna position and antenna type tested (Figure 4 5). The definition of a data point in this max hold erved at each tested frequency. Data analysis All raw data was acquired via the spectrum analyzer in terms of signal strength measured in dBm. Ra dio frequencies being tested were chosen from commercially available products which have to obey RF spectrum r egulations. In the United States, Federal Communications Commission (FCC) regulates operating frequencies and their respective maximum allowed output power (MAOP) Since all three systems do not use the same output powers, antennas, connectors and cables, a calculation had to be done to allow comparison of the systems. Link budget analysis (Shahidi, 1995; Clampitt, 2006) takes into consideration transmitter output power ( P t ) measured signal strength

PAGE 91

91 ( P r ) transmitter antenna gain ( G t ) receiver antenna gai n ( G r ) and various system losses such as connectors, adapters and cables ( L sys ) RF propagation and link budget analysis Radio signal propagation can be analyzed with various models, one of the fundamental RF propagation equations is known as Friis trans mission equation (Friss, 1946), which models line of sight propagation in free space as follow: ( 4 1) Measured signal strength Transmitter output power Transmitter antenna gain Receiver antenna gain Distance of the receiver antenna to the transmitter : Wavelength Wavelength is equal to the speed of light (3x10 8 m/s) over the frequency (in Hz). Therefore, 0 (433MHz) = 0.692m; 0 (915MHz) = 0.327m; 0 (2.45GHz) = 0.125m. According to Friss propagation model (Eq. 4 1), Free Space Path Loss can be calcul ated as (in dB using Log base 10 ) ; ( 4 2) Equation 4 2 shows that received power is equal to the power flow through the effective area of the receiving antenna which is also related to the wavelength and the distance (Friis, 1946). Thi s model (Eq. 4 2) does not account for reflections that are caused by the high metallic environment inside the cargo hold. A better path loss (PL) model (Nikitin and Rao, 2006) would be;

PAGE 92

92 ( 4 3) d: length of the direct ray path : reflection coefficient of the n th reflecting object : length of the n th reflected ray path N: total number of reflections It can be seen from equation 4 3 that reflections have an important impact on path loss calculations. The effects of reflection inside a metal environment was not added to the link budget calculation but was instead being considered as attenuation. So far the propagation model can be improved as follows: (4 4) Where are the syste m losses due to connectors, adapters and cables. Those losses were measured in the lab using an RF signal generator and spectrum analyzer. I n this research attenuation is more specifically defined as path loss and other losses that are not taken into consi deration such as; loss due to pointing error, atmospheric loss, polarization loss and path loss. (4 5) P sys Signal strength data was put in equation 4 5 to calculate the attenuati on. Values of each parameter and calculated attenuation (Eq. 4 5) are shown in Table 4 2 These new attenuation equations were then used to compare results between each test. Data obtained generated a 3 D map of signal attenuation. Full observation of the data can be seen in 12 vertical slices, as shown in the result section in Figures 4 7 to 4 14 or in more details in Appendix B

PAGE 93

93 Link budget analysis also allows comparing data in terms of signal strength. This is useful to evaluate the relationship between signal strength and tag reads. To be able to compare each system to one another, it is optimal to offset the data to simulate the maximum allowed output power (MAOP). In the case of 915MHz and 2.45GHz, the systems used were commercially available readers that should follow FCC regulations, but lab testing still showed deviations. On the other hand, for 433MHz, the RF transmitter used did not have the capability to be set to the MAOP. FCC regulations are written for measured power at a specific range Regul ation for operation in the band 433.5 434.5 MHz states (FCC, 2008) : The field strength of any emissions radiated within the specified frequency band shall not exceed 11,000 microvolts per meter measured at a distance of 3 meters. Regulation for operatio n in the bands 902 928 MHz and 2400 2483.5 MHz states: For frequency hopping systems operating in the 2400 2483.5 MHz band employing at least 75 non overlapping h opping channels : 1 watt. For frequency hopping systems operating in the 902 928 MHz band: 1 w att ( for systems employing at least 50 hopping channels ). The conducted output power limit specified in this section is based on the use of antennas with directional gains that do not exceed 6 dBi. The following calculation provides the MAOP for each freq uency based on the above FCC regulations (4 6 ) In equation (4 6), PD is the power density, E is the electric field strength and o ( impedance of free space) The equation can be re w ritten as: (at r = 3m) (4 7)

PAGE 94

94 The attenuation loss ( Al ) comes from log 10 the received power ( Pr ) equation: where (4 8) Therefore for 433MHz, since the maximum allowed output power measured at 3m is 11m V/m (using log base 10) For 915MHz and 2.45GHz, the limit for 1W at the reader corresponds to 30dBm, plus 6dBi antenna gain, therefore MAOP (at reader) is 36dBm. Each sys MAOP is shown in Table 4 2. The adjustment is the difference between the MAOP and each P sys ). The adjustment was added or subtracted from the data set to mimic an optimal system and permit data comparison. Statistical analysis A mixed linear model was used to test the effect of frequency on attenuation. The main effect of location of the antenna (two antenna positions) was tested on attenuation levels for each frequency and antenna polarization; as well as the effect of antenna polarization ( circular vs. linear ) for each frequency and antenna location with a mixed analysis of variance (ANOVA) model (Littell and Milliken, 2006). The effects of width (three vertical slices), height (three horizontal slices), and depth (12 widthwise slices) of the receiving antenna was also tested on the attenuation level. A residual analysis was performed to check normality and homogeneity of vari ance (Ott and Longnecker, 2004)

PAGE 95

95 All statistical analyses were computed using SAS 9.1 (SAS Inst itute Inc., Cary NC) and significance was accepted at level A more conservative level of acceptance was chosen due to the very large dataset (108 data points per test). Test 2: Validation of Relation between Signal Strength and Tag Reads In an i deal scenario, the measured power levels inside the cargo hold would directly indicate the probability of having successful tag reader communication at a specific point. However, the cargo hold is far from being an ideal space in terms of wave propagation due to the fact that there is interference from metals inside the cargo hold as well as outside sources. Hence, another test was conducted as a proof of concept to show that the measured power levels correspond adequately with a real RFID system performa nce in terms of tag read rates. Signal strength data acquired in Test 1 were used to compare tag readability at 915MHz circular ly and linear ly polarized antenna s Tag readability was tested using 29 AD 210 Gen 1 Class 1 ta gs (Avery Dennison, Flowery Branch, GA) and the ALR 9780 Gen1 915MHz reader (Alien Technology, Morgan Hill, CA). All tags were attached on a sheet of Tyvek material which covered half of the cargo hold cross section as shown in Figure 4 6. The top of the sheet was taped to a s front leg was set longer to allow the sheet to stand as vertical as possible. This set up was moved every meter along the length of the cargo hold on one side (starboard o r port) then it was pivoted (along the central axis of the aircraft) to the other side to gather the other half of the readability data. Data were Technology, Morgan Hill, CA) installed on a laptop c omputer. Read rates obtained correspond to the number of reads per 30 seconds.

PAGE 96

96 Data point comparison Since the signal strength reading s only provide 9 data points per cross section and the readability test use s 5 8 tags the data obtained is averaged for e ach of the 12 slice s. Figure 4 6 shows the location of the 29 tags on the Tyvek sheet which covers half of the cargo hold cross section To obtain data on the other half of the cargo hold, the sheet pivots along the edge that is in the center of the comp artment ( Figure 4 6). Results and Discussion Test 1: Propagation Study All raw data obtained inside the cargo hold (signal level in dBm) were offset by the amount calculated with the link budget as shown in Table 4 2 to permit comparison in terms of attenu ation or signal strength. Comparison of raw measured signal was not appropriate since initial output power was different from one system to the other. Attenuation Attenuation levels for each frequency show how the RF signal fades or attenuates with distan ce. The attenuation levels observed from this experiment are shown in Figures 4 7 to 4 1 4 These graphs are color coded following the visible color spectrum, where red represents high attenuation (70dBm) and purple stands for low attenuation (30dBm). Detai led illustration of e ach slice is shown in Appendix B Comparison of all graphs makes it apparent that 433MHz suffers less attenuation than 915MHz, which also suffers less attenuation than 2.45MHz. As shown earlier in this chapter, lower frequencies have l onger wavelengths, which tend to travel longer distances easier or suffer less path loss. From the graphs it can also be observed that signal variation is more present at lower frequencies than at higher ones. S tandard

PAGE 97

97 deviations for each test are shown in Table 4 3. Despite the higher attenuation, a more uniform dataset is observed at higher frequencies. Effect of frequency on attenuation Figure 4 1 5 shows a simple graph of the distribution of attenuation levels between frequencies for circular antenna tes ts only (and two antenna locations). Linear antenna testing were omitted from the comparison since, because of time constraint, there was no such test done at 433MHz. Figure 4 1 5 shows a proportional relationship between frequency and attenuation. In other words, lower frequencies lead to lower attenuation levels inside the cargo hold. Significant difference between frequencies was statistically tested with a mixed linear ANOVA model (p < 0.0001) ( Appendix C Table C 1). Effect of antenna location and polar ization on attenuation The e ffect of antenna location as well as polarization on attenuation levels for all frequencies mixed up was tested. The results lead to highly significant differences w ith p values <0.0001 (Appendix C Table C 1). However, what is mostly important in reality is the significance of this effect for each frequency and location or polarization. This is explained in the following paragraphs Location The main effect of location of the antenna ( top end vs. ceiling ) was tested on attenuat ion levels for each frequency and antenna polarization with a mixed linear model (Littell and Milliken, 2006). Table 4 4 shows statistical results for each test N one of which is significantly different at = 0.01 level. Polarization The effect of antenn a polarization ( circular vs. linear ) on attenuation for each frequency and antenna location was tested with a mixed linear model (Littell and Milliken, 2006). Again, comparison of antenna polarization was not possible at

PAGE 98

98 433MHz due to the lack of time to p erform linear antenna testing on site. Statistical analysis, as shown in Table 4 5, demonstrates a significant difference between circular and linear polarization at 2.45GHz only. Although statistically different, the mean value of each test differs by 2.5 9dBm, which may or may not have a significant effect on an RFID system. This only depends on how close the signal level is to the sensitivity threshold of the tags Effect of width, height and depth on attenuation The effect of width, height and depth on a ttenuation for each frequency, antenna location and antenna polarization was tested with a mixed linear model (Littell and Milliken, 2006). Width The effect of width was only significant at 2.45GHz for the top end antenna position and linear polarization ( Appendix C Table C 2). Statistical means were as follow s : 62.60dBm (port), 58.39dBm (center) and 61.75dBm (starboard); with p value <0.0001. Difference was significant between center and port as well as center and starboard. Port and starboard were stat istically similar. Height The effect of height was only significant at 915MHz for the top end antenna position and linear polarization ( Appendix C Table C 3). Statistical means were as follow s : 52.66dBm (high), 46.37dBm (middle) and 45.03dBm (low); with p value <0.0001. Difference was significant between high and middle as well as high and low. Middle and low were statistically similar. Depth The effect of depth was significant in all cases except at 2.45GHz for the top end antenna position and linear p olarization ( Appendix C Table C 4). It is interesting to see how antenna polarization affects wave propagation. Just like it is the

PAGE 99

99 case for free space propagation, linear antennas have a longer, but narrower footprint (Dobkin, 2008). This can be observed here by the significant effect of width and height, combined with a non significant effect of depth for linear antenna tests only. In other words, RF signal from a linear antenna attenuates much faster as the receiver moves sideways from the antenna, than it does as the receiver moves away in front of the antenna. Although the width and height results were not significant for all linear antenna tests, p values for linear polarization was lower than for circular polarization in all cases ( Appendix C Tables C 2 and C 3). Signal strength Signal strength comparison allows studying how an actual RFID system would behave if it was i nstalled inside the cargo hold, which will be explained in better detail in test 2 later in this chapter. F or this part of the test the goal is to compare pure signal propagation for each test performed (three frequencies, two antenna positions and one or two antenna polarizations). As mentioned previously, the signal propagation data w ere offset by the amount calculated in the last column of Table 4 2 (adjustment) which brings the dataset to the maximum allowed output power as per FCC regulations. It is important to repeat that regulations for 433MHz are much more strict that those for 915MHz and 2.45GHz. Signal propagation for each test follows the same distribution as for attenuation since both datasets come from the same original data. However, results show a significantly higher signal level at 915MHz compared to the two other frequencies, with averages between 10dBm and 12dBm for 915MHz, compared to between 21dBm and 25dBm for the other two (Table 4 6). This is due to the fact that 915MHz and 2.45GHz start with higher power levels than 433MHz; however, as observed earlier, 915MHz has much lower attenuation levels than 2.45GHz Therefore,

PAGE 100

100 when following FCC regulations, more energy is available i n the cargo hold at 915MHz. G raphical results of each slice can be seen in Appendix D Test 2: Validation of Relation between Signal Strength a nd Tag Reads As previously mentioned, the purpose of this test is to show that there exists a relationship between the measured power levels by the spectrum analyzer and the tag re ad rate by a commercial reader. In an ideal scenario, the power levels would be directly proportional to read rates ex cept irregularities such as saturation at very high power levels. However, the substantial existence of metals inside the cargo hold will result in antenna detuning negatively affecting the read rates. In order to smooth out such effect, both the read rat es and the power levels across the 58 tag reads and 9 signal data points are averaged for each of the 12 cross sectional planes along the cargo hold. This will help understand the relationship between the signal strength and read rates while approximating the proportionality between the two quantities. Two different antennas; circular and linear were used for the experiment with the same reader. The read rates and measured power levels across the 9 data points were averaged for each of the 12 cross section al planes In addition, for better comparison between the power levels and the read rates, each cross sectional power level and read rate average are normalized by the corr esponding global averages. For instance, the global average for power levels in this study including both circular and linear antennas is 13.5dB whereas the global average for read rates across all dimensi ons and antenna types is 10.9. Hence, as Table 4 7 indicates, all the recorded values will be adjusted by the aforementioned averages for improved statistical representation. The first column in table 4 7 is explained as follows: Pc: Average power levels for each cross sectional plane for the circular antenna in dB

PAGE 101

101 Pc adj.: Adjusted average power levels for each cross sectional plane fo r the circular antenna in dB (by 13.5dB) Rc: Average read rates for each cross sectional plane for the circular antenna Rc adj.: Adjusted average read rates for each cross sectional plane for the circular antenna (by 10.9) Pl: Average power levels for each cross sectional plane for the linear antenna in dB Pl adj.: Adjusted average power levels for each cross sectional plane for the linear antenna in dB (by 13.5dB) Rl: Average read rates for each cross sectional plane for the linear antenna Rl adj.: Adjusted average read rates for each cross sectional plane for the linear antenna (by 10.9) Table 4 7 shows that for the cross sections further from the emitting antenna, both the power level and the read rate averages decrease in general confirming the initial assumption that the two values are directly proportional to some extent. Furthermore, figure 4 16 shows the adjusted average power levels and read rates for both circular and linear antennas. In this figure, blue diamonds and red squares show the a verage power levels and read rates for circular and linear antennas respectively across 12 cross sectional planes. Solid lines show the best fitted curves via linear regression with the corresponding R squared values. One can observe from this figure that there are points in the system where higher average power levels do not necessarily correspond to higher average read rates. Heavy concentration of metals around these points and the detuning properties of the metal could have caused such discrepancies as well as the relatively high number of tags being interrogated by the reader resulting in trafficking problems. Nonetheless, the general trends for either antenna as well as the R squared values clearly show the direct proportionality between the two quanti ties.

PAGE 102

102 Another useful observation is to look at the correlation coefficient between the adjusted power level and read rate curves for both circular and linear antennas. In the case of the circular antenna, circular = 0.91, whereas for the linear antenna l inear = 0.96. Although both figures show a strong relationship between average power levels and read rates for either antenna, the relationship is stronger for the lin ear case. This is expected and can be explained by the fact that all the tags were carefu lly placed on the Tyvek sheet in the best possible orientation with respect to the linear emitting antenna, which confirms the fact that linear antennas perform better in use case scenarios where tag orientation is known or controllable. The results for b oth tests show that, even though there are many factors to be considered when estimating a commercial RFID reader performance in a given environment, measuring the signal strength at key locations in the application space would indicate the weak and strong points in the system. It is well known that each system built by different manufacturers will have technical differences in terms of sensitivity, coding technology, etc. However, this experiment indicates that the power measurements explained in this chap ter which include the three frequencies of 433MHz, 915MHz and 2.4GHz can all serve as a guideline when determining which system would perform better in terms of RFID tag read rates under the same circumstances and technical specifications. Conclusion This test demonstrate d that frequencies have a major influence on signal propagation, especially inside a metal environment. Lower frequencies suffer less attenuation over distance, but have higher variation within the cargo hold. It was also shown that antenna location did not deliver significantly different results. However,

PAGE 103

103 antenna polarization can have a significant effect on signal propagation in some cases, and therefore should not be omitted when designing an RFID system for air cargo transportation. More ove r, FCC regulations restrict output powers at 433MHz more than at 915MHz and 2.45GHz, which lead s to the conclusion that more RF signal is available in the cargo hold at 915MHz It was also demonstrated that the relationship between signal strength and t ag reads is an important factor to take into account when considering the installation of an RFID system inside an application space with nonzero interference.

PAGE 104

104 Table 4 1. Specifications of the three RF systems used. Frequency RF system Antenna Type Pol arization Gain Model & Manufacturer 433 MHz Chipcon CC1100 RF transmitter (Texas Instruments Inc., Dallas, TX) Emitter Circular 9 dBi SPA 430, Huber + Suhner AG, Essex, VT Receiver Linear (omni) 0 dBi B 368 1, How Tsen Intl. Electronics Metal Co.,Ltd. Shi n Wu Hsiang, Tao Yuan Hsien, Taiwan 915 MHz 915 MHz Alien RFID reader ALR 9780 (Alien Technology, Morgan Hill, CA); Emitter Circular 6 dBi ALR 9610 BC, Alien Technology, Morgan Hill, CA Emitter Linear 5.9 dBi ARL 9610 AL, Alien Technology, Morgan Hill, C A Receiver Linear (omni) 2.5 dBi EXR902TN, Laird Technologies, Schaumburg, IL 2.45 GHz 2.45 GHz Alien RFID reader ALB 2484 (Alien Technology, Morgan Hill, CA) Emitter Circular 6 dBi 2AC 001, Alien Technology, Morgan Hill, CA Emitter Linear 15 dBi Com 24015p, Antenna World, Miami, FL Receiver Linear (omni) 8 dBi MRN 24008SM3, Antenna World, Miami, FL Table 4 2. Calculated parameters for the attenuation equation (Eq. 4 5) and maximum allowed output power adjustment Systems P t (dBm) G t (dBi) G r ( dBi) L sys (dBm) Attenuation P sys P r (dBm) MAOP (dBm) Adjustement for MAOP (dBm) 433MHz circular 10.34 9.0 0.0 2.50 16.85 P r 20.4 +3.55 915MHz circular 34.97 6.0 2.5 3.95 39.52 P r 36 3.52 915MHz linear 34.97 5.9 2.5 3.95 39.42 P r 36 3.42 2 .45GHz circular 32.24 6.0 8.0 7.26 38.99 P r 36 2.99 2.45GHz linear 32.24 15.0 8.0 6.71 48.54 P r 36 12.54 Table 4 3. Averages and standard deviations of attenuation levels for each test. 433MHz 915MHz 2.45GHz circular circular linear circular l inear top end ceiling top end ceiling top end top end ceiling top end Average 44.20 44.17 48.00 46.91 48.02 58.32 58.79 60.91 Std dev. 4.75 4.72 3.53 3.39 5.72 3.19 3.06 4.16

PAGE 105

105 Table 4 4. Statistical analysis results for the effect of antenna location Frequency Polarization Location Mean F value p value 433 MHz Circular Top End 41.70 0.00 0.9615 Ceiling 41.67 915 MHz Circular Top End 48.00 5.27 0.0227 Ceiling 46.91 2.45GHz Circular Top End 58.32 1.20 0.2744 Ceiling 58.79 Table 4 5. Statistical analysis results for the effect of antenna polarization. Frequency Location Polarization Mean F value p value 433 MHz Top End Circular N/A N/A N/A Linear N/A 915 MHz Top End Circular 48.00 0.00 0.9683 Linear 48.02 2.45G Hz Top End Cir cular 58.32 26.34 < 0.0001 Linear 60.91 Table 4 6. Signal strength data for each test, averaged per vertical slice, and total cargo hold (Avg). 433MHz 915MHz 2.45GHz Slices circular circular linear circular linear (m) top end ceiling top end c eiling top end top end ceiling top end 1 aver age per slice (dBm) 16.2 26.6 7.6 14.8 6.2 18.6 24.0 24.0 2 18.6 23.7 9.5 13.5 8.5 19.4 23.6 24.7 3 17.3 20.1 10.0 10.8 8.9 20.8 22.6 24.8 4 19.3 19.6 10.5 11.3 9.0 22.4 21. 6 23.8 5 18.8 20.4 11.2 8.7 9.4 21.6 20.3 24.7 6 21.9 16.5 11.7 7.3 11.3 21.8 18.8 25.1 7 23.1 18.6 11.9 6.4 13.1 22.3 19.9 23.1 8 22.9 19.7 12.7 9.2 13.6 23.4 22.4 25.4 9 24.0 22.6 12.3 10.9 16.0 23.1 23.2 24.6 10 23.1 22.9 15.0 11.4 15.6 24.0 25.2 25.6 11 23.8 22.2 16.7 12.6 16.5 24.7 25.8 26.9 12 26.7 22.5 14.9 14.3 15.9 25.9 26.1 26.2 avg 21.3 21.3 12.0 10.9 12.0 22.3 22.8 24.9

PAGE 106

106 Table 4 7. Table summarizing the recorded and adjusted power levels (P) and read rates (R) for circular and linear antennas across the 12 cross sectional planes. Sectional planes 1 2 3 4 5 6 7 8 9 10 11 12 P c 10.6 12.5 13.0 13.5 14.2 14.7 14.9 15.7 15.3 18.0 19.7 17.9 P c adj. 2.9 1.0 0.5 0.0 0.7 1.2 1.4 2.2 1.8 4.5 6.2 4.4 Rc 18.6 14.7 16.2 14.3 13.2 9.6 4.6 5.1 1.3 0.7 0.4 0.0 Rc adj. 7.7 3.8 5.3 3.4 2.3 1.3 6.3 5.8 9.6 10.2 10.5 10.9 P l 6.2 8.5 8.9 9.0 9.4 11.3 13.1 13.6 16.0 15.6 16.5 15.9 P l adj. 7.3 5.0 4.6 4.5 4.1 2.2 0.4 0.1 2.5 2.1 3.0 2.4 Rl 20.1 19.8 20.9 16.0 18.4 13.8 11.7 11.0 8.4 7.0 6.1 10.0 Rl adj 9.2 8.9 10.0 5.1 7.5 2.9 0.8 0.1 2.5 3.9 4.8 0.9 Figure 4 1 Section of an aircraft fuselage (Airbus A380)

PAGE 107

107 Figur e 4 2 DC 10 30F from Arrow Cargo. Figure 4 3. Typical fuselage section of a DC 10 30F, lower cargo hold circled in blue (Boeing 2010c)

PAGE 108

108 Figure 4 4. Cargo hold dimensions and RF emitting antenna positions. Figure 4 5. Data point positions in the 3x3 grid. Twelve 3x3 grids are measured long the length of the cargo hold, every meter.

PAGE 109

109 Figure 4 6. Tag readability test configuration. Tyvek sheet with 29 RFID tags (circled) covering half of the cargo hold cross sec tion Figure 4 7. Attenuation surface plots for each vertical slice of 3x3 data point at 433MHz, circular antenna and top end antenna position.

PAGE 110

110 Figure 4 8. Attenuation surface plots for each vertical slice of 3x3 data point at 433MHz, circular antenna and center ceiling antenna position. Figure 4 9. Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz, circular antenna and top end antenna position. Figure 4 10. Attenuation surface plots for each vertical slice of 3x3 data p oint at 915MHz, linear antenna and top end antenna position.

PAGE 111

111 Figure 4 11. Attenuation surface plots for each vertical slice of 3x3 data point at 915MHz, circular antenna and center ceiling antenna position. Figure 4 12. Attenuation surface plots for ea ch vertical slice of 3x3 data point at 2.45GHz, circular antenna and top end antenna position. Figure 4 13. Attenuation surface plots for each vertical slice of 3x3 data point at 2.45GHz, linear antenna and top end antenna position.

PAGE 112

112 Figure 4 14. Attenu ation surface plots for each vertical slice of 3x3 data point at 2.45GHz, circular antenna and center ceiling antenna position.

PAGE 113

113 Figure 4 1 5 Distribution (in percentage) of each frequency tested, for circular antenna only and two antenna locations.

PAGE 114

114 F igure 4 1 6 Comparison of the change in average power levels and tag read rate s for both antennas through linear regression

PAGE 115

115 CHAPTER 5 RADIO FREQUENCY INTE RACTIONS WITH AIR CA RGO CONTAINER MATERI ALS FOR REAL TIME MONITORING Introduction Products, such as food, pharmaceuticals and flowers, are at high risk of perishing from various adversities along the cold chain. The parties involved should control when possible, and at the very least monitor the conditions of the goods in order to ensure their quality a nd to comply with all legal requirements. Among environmental parameters during transport, temperature is the most important in maintaining the shelf life of the products (Nunes et al., 2006; Zweig, 2006; Jedermann et al., 2009). Cold chain globalization, there is a growing need for fresh products to be delivered year round all over the world, thus, temperature sensitive items are likely to be shipped by air because of their relatively short shelf life. Unfortunately, a faster transit time do es not always imply controlled temperature throughout transportation. In contrast, during airport operations, loading, unloading, air transportation or warehouse storage, perishable goods often suffer from temperature abuse either due to difficulties in co ntrolling the temperature, absence of refrigerated facilities, or lack of information about produce characteristics and needs (Nunes et al., 2003). O f approximately 2.6 million tons of perishables air freighted in 2008, nearly 30% is estimated to be lost d ue to handling and temperature abuse (Catto Smith, 2006). In a previous study, mond et al. (1999) showed that the environmental conditions during airport operations could, in fact, be very far from the optimum for fruits and vegetables. Moreover, in a str awberry quality study, Nunes et al. (2003) showed that greater losses in quality occurred during simulation of the airport handling operations, in flight, and retail display than during

PAGE 116

116 warehouse storage at the grower, truck transportation to or from the a irport, or during backroom storage at the supermarket. Temperature monitoring Currently, most digital temperature loggers have to be connected to a host device to download data, and as a result, have limited real time data interactivity, which result s in after the fact analysis for claims, loss in quality and related issues. Radio frequency identification (RFID) temperature loggers function wirelessly which allows for real time information transfer. Active or semi passive RFID tags can support one or many sensors as well as the unique ID that RFID technology provides by design. The RFID tag, with associated hardware and software gives the added benefit of having the item scanned on receipt, so that if an alert (alert triggers are programmable prior to shipp ing) is active, the receiver knows immediately (not after the fact) that there is a potential problem with the shipment and can spend the time required on specific shipments rather than going through random inspections (Jedermann et a l., 2007). Many studie s have shown the effectiveness of RFID in monitoring product temperature during transit ( mond 2007; Jedermann and Lang, 2007; Jedermann et al., 2007; Ketzenberg and Bloemhof Ruwaard, 2009). RFID technology or many cases, the technology is not flawless. Certain materials, like metals and water based liquids, are challenging for RFID systems (Foster and Burberry, 1999; mond 2008) and are generally referred to as being RF Opaque. The behaviour of radio freque ncy around metal has been studied extensively (Dobkin and Weigand, 2005; Griffin et al., 2006; Prothro et al., 2006; Sydanheimo et al., 2006). Because aluminum is a very good conductor (conductivity 38 MS/m), an incident electromagnetic wave totally reflec ts from

PAGE 117

117 the metallic surface with a phase reversal (Cheng, 1993; Reitz et al., 1993). Moreover, metallic surface of the object in the vicinity of an antenna changes its radiation pattern, input impedance, radiation efficiency and resonant frequency. These changes depend on the size and shape of the metallic object and also on the distance of the antenna from the object (Raumonen, 2003; Mo and Zhang, 2007). Mo and Zhang (2007) also demonstrated that RFID tags placed 1/4 wavelength away from the metallic surf ace enhances the readability of the tags. Not only metallic materials, but also dielectrics (or electrical insulators) cause reflections. Other m aterials affect part of the incident energy and transmit the rest. The exact amount of transmission and reflect ion is also dependant on the angle of incidence, material thickness, and dielectric properties (Blaunstein and Christodoulou, 2007). On the other hand, little or no reflection occurs when electromagnetic waves penetrate directly through objects such as pap er, non conductive plastics or textiles (Penttil et al., 2006). These materials, including most composites, are non absorbing and possess low refractive indexes. Such materials are generally referred to as being RF lucent. Air Cargo While the world is ta lking about climate change, the airline industry is looking at ways to be more fuel efficient to minimize their operational costs as well as their impact on the environment. One way to do so is to reduce the weight, and minimizing weight without compromisi ng the business volume is feasible by using lighter containers, or Unit Load Devices ( ULDs ). Composite ULDs can save up to 25% of the tare weight of a traditional aluminum ULD (Nordisk, 2010). For illustration: A Boeing 747 400 aircraft, equipp ed with 16 s tandard aluminum ULD s normally has a total of

PAGE 118

118 1216kg empty container weight. Alternatively, by using ultra light composite ULDs the combined empty container weight total would be approximately 880kg. Furthermore, composites containers are easier to repair and require fewer visits to a repair station than aluminum units (Saunders, 2003). Kevlar ULDs are constantly replacing older Aluminum ULDs still add up to 43% of their fleet, whereas Lexan containers count for the remaining 18%. Considering that ULDs have an approximate usable life of 10 years, Aluminum ULDs will most likely be outnumbered by composite containers relatively quickly. This study focuses on the air transp ortation part of the cold chain. RFID is not yet a widespread technology in the transportation industry, but its potential value makes it worth the investigation effort. The objective of this study is to explore the possibility of real time temperature mon itoring during air cargo operations by researching the effect of container wall materials on RF propagation. Five different ULD materials were chosen for this study: Aluminum, Duralite, Herculite, Kevlar and Lexan. Due to the fact that the RF behavior of materials depend s on size, shape and thickness, all samples used for this study were collected from an airline container maintenance facility and therefore represent the true properties for each material. Initial hypotheses are that only Aluminum samples will not allow RF transmission, whereas all other materials will transmit radio waves with negligible interference. Materials and Methods Three radio frequencies (433MHz, 915MHz and 2.45GHz) were tested against five different air cargo materials as describ ed in the introduction: Aluminum, Duralite, Herculite, Kevlar and Lexan. Duralite is a thick fibreglass woven composite. Herculite

PAGE 119

119 (or Twintex P PP) is a thermoplastic glass reinforcement panel made of commingled E Glass and thermoplastic filaments. Kev lar is made with high strength para aramid fiber and Lexan is a translucent polycarbonate. For the first two tests, each sample was a square sheet of 0.305m long sides and thicknesses of 1.00, 1.80, 1.00, 0.50 and 1.80mm respectively. For the third test both samples were squares of sides 1.22m long. This series of tests was performed inside an anechoic chamber of dimensions 2.05m high, 1.90m wide and 2.70m deep. The wall materials were Eccosorb VHP 12 NRL and Eccosorb FS 100 NRL (Emerson & Cuming Microwav e Products N.V., Westerlo, Belgium), a solid, pyramidal shaped, carbon loaded urethane foam absorber. Each frequency was generated by an RF signal generator (Agilent N9310A, Agilent Technologies, Santa Clara, CA); power supply (XTR 33 25, Xantrex technolog y, Burnaby, BC, Canada); and power amplifiers (5803039A and 5303081, Ophir RF, Los Angeles, CA). This equipment was located outside of the anechoic chamber during testing. The RF output of this system was conveyed to the anechoic chamber via a 50m long LMR 400 low loss cable. Each frequency was tested with a particular set of emitter and receiver antennas (Table 5 1) and only one frequency was tested at a time. Three tests were administered to determine the effects of the materials on RF propagation. For al l tests, the received signal was measured with a spectrum analyzer (RSA3303B, Tektronix, Beaverton, OR), also kept outside the door of the anechoic chamber during testing. The definition of a data point in this experiment is a 200 sample average of the pea k signal power observed at each tested frequency. One frequency was tested at a time and all data were analyzed with reference to the control data point (no material sample present).

PAGE 120

120 Test 1 The goal of this test was to quantify the reflection and absorpti on characteristics of each material. Inside the chamber, the emitting antenna, receiver antenna and material samples were arranged in a row on a Plexiglas table with a Styrofoam plate to help hold everything in place (Figure 5 1). The table was centered in the room 0.38m above the floor, just over the anechoic chamber surface material responsible from absorbing outside RF radiation. The emitting antenna was positioned vertically, beaming towards the back of the room. The receiving antenna was also positione d vertically with specific intervals based on the radiation wavelength (at /2, 3 /4, /4 or /2), and the material sample was positioned at in front of the emitting antenna. Wavelength in meters is calculated as; (5 1) Where c is the speed of light in m/s and f is the frequency in Hz. In other words, the sample was 0.692m from the 433MHz antenna; 0.328m away from the 915MHz antenna; or 0.125m away from the 2.45GHz antenna. The respective receiver antennas were consecutively placed and wavelengths away from the sample, on both sides (Figure 5 1). Ideally, test 1 should have been accomplished with infinite planes of each sample. Reality is different, and material availability was limiting. This design is interesting in the way that it procures information on more aspects of radio frequency beh avior, such as wave scattering and diffraction around sharp obstacles. In reality, those effects exist and are inevitable components in RFID applications. Test 2 In order to achieve a more uniform dataset, the goal of this test was to isolate the receiver antenna from the knife edge diffraction effect. The sample was framed with a

PAGE 121

121 solid, pyramidal shaped carbon loaded urethane foam absorber (anechoic chamber wall material). The foam pyramids were glued onto a 0.05m thick Styrofoam sheet and were positione d to leave the center part of the 0.305m by 0.305m square empty to place the samples as shown in Figure 5 2A. The samples were again positioned one wavelength away from the emitting antennas, and the receiver antenna was taped behind the sample (Figure 5 2 B). Three repetitions of each data point were performed for statistical analysis. Statistical analysis consisted of one way ANOVA to show significant differences between the materials for each frequency. Multiple comparisons of means were performed with Bo nferroni adjustments. All statistical analyses were computed using SAS 9.1 (SAS Institute Inc. 2003). Test 3 Following the thought process from test 1 to test 2, it was determined that an additional test was needed to clarify the effect of using a smalle r sample and therefore show more realistic properties of RF lucent and RF opaque materials. Since larger samples were not available in all materials, Aluminum and Kevlar were chosen and samples of 1.22m x 1.22m were tested (which is 16 times larger than the previous sample size). Those two materials were chosen because of availability, but also because of their wide use in the air cargo container fleet. Aluminum and Kevlar Moreover, those two materials can represent typical RF lucent and RF opaque materials encountered in the air cargo industry. The same set up as test 1 was used, which means that the emitting antenna was positioned vertically, beaming towards the back of t he room. T he receiving antenna was also positioned vertically with specific intervals based on the radiation wavelength (at

PAGE 122

122 /2, 3 /4, /4 or /2); and the material sample was positioned at in front of the emitting antenna. Except this time because o f sample size, the antennas and samples were hung from the ceiling of the chamber with strings, instead of being held in place on the table. Three repetitions of each data point were performed for statistical analysis Results and Discussion Test 1 The res ults showed a very strong effect for Aluminum on RF transmission, and minimal interaction for all other sample materials. All comparisons were made between the control and each sample. Table 5 2 shows values obtained for the control measurements (no materi al present), whereas Table 5 3 illustrates the signal deviations from the control (control subtracted from each signal strength measurement). Receiver antenna positions are measured from the emitting antenna and sample materials are positioned at 433MHz Results show weaker signal levels in front of the Aluminum sample at /2 ( 1.52dBm) and higher signal strength at 3 /4 (+2.39dBm) (Table 2 3). This confirms the observation made by Mo and Zhang (2007), which is when an electromagnetic wave hits a metalli c surface, it reflects with a 180 phase reversal. This causes signal cancellation at /2 and signal amplification at /4 from the metallic surface. In our case, /4 from the Aluminum sample is 3 /4 distance from the emitting antenna. All other samples sho w no considerable loss or gain from reflections when the receiver antenna was positioned in front of the samples (within 0.09dBm from control). As far as signal transmission through the samples, it is understandable that only the Aluminum sample

PAGE 123

123 offers co nsiderable signal blocking, with signal loss of 5.45dBm at + /4 and 2.70dBm at + /2. All other materials were within 0.50dBm from the control. 915MHz In this part of the experiment, signal strength i n front of the Aluminum sample was increased in both /2 and 3 / 4 cases, although the increase was greater at 3 /4 (+7.11 vs. +3.98). This could be caused by signal scattering since the plate size (0.305m) wa s slightly smaller but very close to the wavelength at 915MHz (0.325m). Since for the case of 915MHz the wavelength and the dimensions of the material (obstacl e) we r e of similar sizes, the set up wa s in the resonance range (Finkenzeller, 2003). Therefore, the behavior of RF radiation may not follow traditional rules such as the one stated by Mo and Zhang due to the unpredictable nature of edge diffractions (Long hurst, 1967 ) as well as resonance. Signal wa s also slightly reflected from of other materials, Lexan being the second most reflecting with +1.18dBm gain. In the case of signal transmission, similar results we re observed as with 433MHz, except the signal l oss is greater, with 19.70 and 11.74 at + /4 and + /2 res pectively. All other materials we re within 0.17dBm of the control. 2.45GHz reflection with a loss of 6.86dBm and a gain of +2.81dBm at /2 and 3 /4 respective ly. All other materials we re within 0.58dBm of the control. Moreover, the signal loss behind the samples was obvious with 37.99dBm at + /4 and 34.37dBm at + /2, all other materials being within 0.55dBm of th e control.

PAGE 124

124 Looking at signal transmission behind the Aluminum sample, it was noticeable that the signal loss increases with the frequency. This was caused by the ratio of the wavelengths and the materials sample size. At 433MHz, the wavelength was more tha n two times longer than the sample size (0.692m and 0.305m respectively ); at 915MHz, both dimensions wer e similar ( = 0.325m); and at 2.45GHz, the wavelength was about half of the sample size ( = 0.125m). When a radio wave impinges an obstacle larger than its wavelength, reflection occurs. However, when a wave hits an obstacle smaller than its wavelength, scattering occurs and wave patterns are redirected with random phase and amplitude (Blaunstein and Christodoulou, 2007). It is also noticeable that there was minor signal amplification behind the non metallic samples at 433MHz and 915MHz. This can be explained by th e fact that the sample size was smaller than the wavelengths, which allows waves to travel around the edge diffraction and explains the redirection of electromagnetic waves when they hit a solid obstacle such as the edge of the material sample in this experiment (Kumar et al., 2007). Knife edge diffraction is described by Huygens Fresnel principle which states that such an obstruction (the edge of the material in this case) will act as a secondary source o f RF radiation (Longhurst, 1967). Depending on the wavelength of the electromagnetic signal, the effe cts of this secondary source could be observed at different points in the measurement field, in this case, amplification behind the non metallic samples, however, the discussion of this phenomenon in greater detail is beyond the scope of this text.

PAGE 125

125 Test 2 When six treatments (material samples) we re compared, all results are reported as significant when P < 0.05 and the Aluminum sample wa s the only one sig nificantly different from the others for all three frequencies. Due to the nature of the second experiment it would be expected to obtain higher attenuation at lower frequencies b ecause shorter wavelengths travel more easily inside the open frame within th e foam absorber material. However, one should note that this observation is affected by two important parameters: the electromagnetic properties of the container samples as well as the absorption profile of the urethane foam absorber, which is proportional to the frequency (Eccosorb, 2008). For instance, for free air (control) the signal strength at 433MHz is 3.83dBm whereas the signal strength at 915MHz is 9.53dBm. This clearly shows the attenuation from the wavelength dimension at lower frequency as expe cted. However at 2.45GHz, the signal power wa s attenuated to 6.96dBm, which is explained by the fact that the foam absorber material has higher absorption coefficients at higher frequencies. Test 3 Table 5 5 shows values obtained for the control measureme nts (no material present), whereas Table 5 6 illustrates the signal deviations from the control (control subtracted from each signal strength measurement). All comparisons were made between the control measurement and each sample. The results showed a very strong effect for Aluminum on RF transmission, and minimal interaction for Kevlar as expected from the previous tests. This test, however, show ed a much more important attenuation level behind the Aluminum sample when comparing to the results obtained in test 1 (same test set up, different sample sizes) for frequencies 433MHz and

PAGE 126

126 915MHz. As stated previously, 433MHz corresponds to a wavelength of 0.692m, whereas 915MHz corresponds to 0.328m and 2.45GHz to 0.125m. As opposed to the earlier tests, the dim ensions for the samples in test 3 were 1.22m x 1.22m, which are larger than all three wavelengths in this case. Similar to the previous tests, the attenuation levels measur ed behind the Aluminum sample wer e still proportional to the frequency tested. Atte nuation at 433MHz is lower than attenuation at 915MHz and 2.45GHz, but the divergence wa s not as strong as in test 1 which can be explained by the use of a larger sample between the emitting and receiving antennas. The general trend shows that an infinite aluminum plane would probably lead to similar results in terms of total attenuation for all frequencies. As far as reflection, a similar pattern wa s observed where the RF waves that reflects from the aluminum surface increase the signal level in front of t he sample in all cases. This confirms again what Mo and Zhang (2007) had previously observed, which is when the electromagnetic wave reflects off a metallic surface, it causes signal cancellation at /2 and signal amplification at /4 from the metallic surface. As previously stated, the goal of this test was to show that a larger than wavelength sample size would affect the signal level measurements behind the samples by eliminating secondary effects such as edge reflection discussed in the previous secti ons and uniformize the results. The values presented in tables 5 5 and 5 6 show that this goal was accomplished by measuring higher power levels in front of the sample material for Aluminum and behin d t he sample material for Kevlar.

PAGE 127

127 Conclusion This test demonstrated the effects of five commonly used air cargo container wall materials on RF propagation at three different frequencies. The reflection and absorption characteristics of each material were quantified. Three different tests were utilized to analyze the characteristics of RF propagation in greater detail for each material and the results from all experiments showed a very strong effect of Aluminum on RF transmission and minimal interaction fo r all other sample materials as expected. These finding s suggest that the use of non metallic containers for air transportation of perishable products should make real time temperature monitoring possible by allowing RF waves to transmit through the wall s urface effortlessly. This goes well with the current trend that encourages the use of Kevlar containers over aluminum ones because of their much lighter tare weight.

PAGE 128

128 Table 5 1. Specifications of the six antennas used. Frequency Antenna Polarization Gai n Model & Manufacturer 433 MHz Emitter Circular 9 dBi SPA 430, Huber + Suhner AG, Essex, VT Receiver Linear (omni) 0 dBi B 368 1, How Tsen Intl. Electronics Metal Co.,Ltd. Shin Wu Hsiang, Tao Yuan Hsien, Taiwan 915 MHz Emitter Circular 8 dBi SPA 915, Hub er + Suhner AG, Essex, VT Receiver Linear (omni) 2.5 dBi EXR902TN, Laird Technologies, Schaumburg, IL 2.45 GHz Emitter Circular 6 dBi 2AC 001, Alien Technology, Morgan Hill, CA Receiver Linear (omni) 8 dBi MRN 24008SM3, AntennaWorld, Miami, FL Table 5 2. Signal strength measurements (dBm) for control (no sample), test 1. Receiver antenna positions are measured from the emitting antenna. Frequenc ies Receiver antenna position s /2 3 /4 + /4 + /2 433MHz 11.07 9.25 8.65 1.80 915MHz 13.90 11.03 7.97 6.15 2.45GHz 8.94 7.85 6.64 6.12 Table 5 3 Signal strength deviation (dBm) from control (no sample) for test 1. Receiver antenna positions are measured from the emitting antenn a and sampl e materials are positioned at Frequenc ies Materials Receiver antenna position s /2 3 /4 + /4 + /2 433 MHz Aluminum 1.52 2.39 5.45 2.70 Duralite 0.02 0.09 0.02 0.18 Herculite 0.03 0.02 0.15 0.50 Kevlar 0.02 0.02 0.09 0.40 Le xan 0.01 0.02 0.08 0.14 915 MHz Aluminum 3.98 7.11 19.70 11.74 Duralite 1.09 0.40 0.17 0.12 Herculite 0.84 0.30 0.08 0.03 Kevlar 0.96 0.32 0.09 0.00 Lexan 1.18 0.16 0.17 0.10 2.45 GHz Aluminum 6.86 2.81 37.99 34.37 Duralite 0.22 0.38 0. 32 0.10 Herculite 0.17 0.54 0.51 0.50 Kevlar 0.15 0.30 0.55 0.16 Lexan 0.07 0.58 0.29 0.23

PAGE 129

129 Table 5 4. Signal strength measurements (meanSD) (dBm) for control, plus signal strength deviation between material samples and control at thre e frequencies for test 2 (n=3). Materials Frequencies 433MHz 915MHz 2.45GHz Control 3.83 0.05 9.53 0.05 6.96 0.01 Aluminum 15.16 0.06 20.49 0.18 35.47 0.23 Duralite 0.05 0.03 0.21 0.01 0.19 0.03 Herculite 0.04 0.03 0.29 0.02 0.33 0.01 Kevla r 0.01 0.01 0.35 0.01 0.37 0.01 Lexan 0.06 0.04 0.53 0.65 0.07 0.03 Table 5 5 Signal strength measurements (dBm) for control (no sample), test 3 Receiver antenna positions are measured from the emitting antenna. Frequenc ies Receiver antenna po sition s /2 3 /4 + /4 + /2 433MHz 10.60 8.50 4.53 1.97 915MHz 14.93 11.72 7.76 6.41 2.45GHz 11.04 8.97 5.47 4.79 Table 5 6. Signal strength deviation (dBm) from control (no sample) for test 3 Receiver antenna positions are measured from the emit ting antenna and sampl e materials are positioned at Frequenc ies Materials Receiver antenna position s /2 3 /4 + /4 + /2 433MHz Aluminum 0.36 3.18 20.01 15.33 Kevlar 0.74 0.30 1.48 1.27 915MHz Aluminum 2.17 6.20 26.84 24.53 Kevlar 0.19 0.42 0.13 0.11 2.45GHz Aluminum 0.84 3.20 38.12 37.48 Kevlar 0.23 0.40 0.00 0.56

PAGE 130

130 Figure 5 1. Diagram of the anechoic chamber setup for test 1. Note that four receiver antennas are shown for illustrative purposes as only one receiver antenna i s used at a time for each test. Figure 5 2. Anecho ic chamber set up for test 2. A ) The sample material is surrounded by pyramidal shaped, carbon loaded urethane foam absorber and is placed one wavelength from the emitting antenna. B ) The receiver ant enna is taped behind the material sample. A B

PAGE 131

131 CHAPTER 6 TEMPERATURE MAPPING INSIDE AIR CARGO CON TAINERS DURING AIRSIDE OPERATIONS Introduction Temperature is well regulated in the cabin of most passenger flights, but it is not necessarily the case insid e the cargo hold or in freighter flights. Temperature distribution and variability in cargo holds depends on many factors such as weather (air temperature, wind speed, sun radiation), duration of flight, type of aircraft (ability to control cargo ambient t emperature), altitude, and transit time on the tarmac. Aircraft Temperature control inside the cargo hold can be regulated at different levels. On one end, the only heat available in the cargo hold of some airc raft comes from leaks in the cabin floor; som e other airplanes have ventilation systems that can re circulate the cabin air into the cargo hold; and the more sophisticated ones have their own heating systems designed to keep the cargo from freezing ( Air Canada, 2007 ). Typically, larger and newer airc rafts have more options. When aircrafts are categorized by size, two major categories emerge, which are narrow body and wide body aircrafts. Narrow body aircrafts have one aisle and two rows of seats in the cabin, whereas wide body aircrafts have two aisle s and three rows of seats in the cabin floor. As a consequence, the wide body aircraft has a larger cargo compartment and can also be used on longer routes. Studies on temperature profiles inside cargo holds state many different temperature ranges. Syverse n et al. (2008) found that 49.5% of shipments were exposed to high temperatures (greater than 29.4 C), 14.6% to low temperatures (less than 7.2 C), and 61% to temperature variations of 11 C or more. It was also shown that temperature depends on ULD posi tion inside the cargo hold as well as which cargo hold is used ( mond et al., 1999 ).

PAGE 132

132 Air cargo operations All cargo being planned on a flight is built up onto unit load devices ( ULDs ) or in tub carts (for bulk) a specific time prior to flight departure. T his specific time depends on the airport, destination (domestic vs. international), size of freight, and type of product (priority vs. general freight) (T. Howard, 2010, personal communication). The freight usually leave s the warehouse ( on the airside) for a maximum of 2h before flight time, but typically spends between 45 and 90 minutes on the tarmac before being loaded onto the aircraft (T. Howard, 2010, personal communication ). In practice, if something happens to delay the loading activity of the plane, the freight c an be left on the tarmac for extended periods of time, regardless of fluctuations while waiting on the tarmac, for example, sun radiation, air temperature, ULD wall materia l properties, wind speed and direction, etc ( Villeneuve et al., 2001 ). Perishables. Products, such as food, pharmaceuticals or flowers, are at high risk of perishing from various adversities along the cold chain. Among environmental parameters during trans port, temperature is the most important in maintaining the shelf life of the products (Nunes et al., 2006; Villeneuve, 2006 ; Zweig, 2006; Jedermann et delivered year r ound all over the world, Temperature sensitive items are likely to be shipped by air because of their relatively short shelf life. Unfortunately, a faster transit time does not always imply controlled temperature throughout transportation. In contrast, dur ing airport operations, loading, unloading, air transportation or warehouse storage, perishable goods often suffer from temperature abuse either due to difficulties in controlling the temperature, absence of refrigerated facilities, or lack of information

PAGE 133

133 about produce characteristics and needs (Nunes et al., 2003). On approximately 2.6 million metric tons of perishables air freighted in 2008, nearly 30% is estimated to be lost due to handling and temperature abuse (Catto Smith, 2006). In a previous study, mond et al. (1999) showed that the environmental conditions during airport operations could, in fact, be very far from the optimum for fruits and vegetables. Moreover, in a strawberry quality study, Nunes et al. (2003) showed that greater losses in qualit y occurred during simulation of the airport handling operations, in flight, and retail display than during warehouse storage at the grower, truck transportation to or from the airport, or during backroom storage at the supermarket. Many more studies showed that important temperature fluctuations can occur during airport ground operations ( Bollen et al. 1998; Villeneuve et al., 2000; Vill eneuve et al., 2001 ). Moreover, because perishables are mostly season dependant, transportation companies do not offer sp ecial treatments like they would if they were available year round (Villeneuve, 2006). Temperature monitoring Currently, most digital temperature loggers have to be connected to a host device to download data, and as a result, have limited real time data interactivity, which result in after the fact analysis for claims, loss in quality and related issues. Radio frequency identification (RFID) temperature loggers function wirelessly which allows for real time information transfer (Rao, 1999, Lahiri, 2006). Active or semi passive RFID tags can support one or many sensors as well as the unique ID that RFID technology provides by design. The RFID tag, with associated hardware and software will add the benefit of having the item scanned on receipt, so that if a n alert (alert triggers are programmable prior to shipping) is active, the receiver knows immediately (not after the fact) that there is a potential problem with the

PAGE 134

134 shipment and can spend the time required on specific shipments rather than going through r andom inspections (Jedermann et al., 2007). Many studies have already shown the effectiveness of RFID in monitoring product temperature during transit ( mond 2007; Jedermann and Lang, 2007; Jedermann et al., 2007; Ketzenberg and Bloemhof Ruwaard, 2009). S ince temperature control inside cargo holds is a weak link in the air cargo cold chain, the objective of this study is to evaluate the temperature distribution inside the ULDs during airside operations Comparison will be made between long and short flight s, as well as between wide and narrow body aircrafts. Hypotheses are that temperature will drop lower at the bottom part of containers as well as during longer flights in general Materials and Methods In June 2010, 12 ULDs were shipped on 10 different on e way flights (Table 6 1). The number of ULDs monitored per aircraft was limited by the amount of cargo available to travel on that route that specific day. All flights were managed by Air Canada and originated from Toronto, Canada (YYZ). Three ULDs travel ed to Montreal, Canada (YUL); one ULD flew to Vancouver, Canada (YVR); four ULDs flew to London, UK (LHR); and four ULDs were shipped to Frankfurt, Germany (FRA). All these ULDs were reloaded at destination to be shipped back on a returning flight. All con tainers shipped on Canadian routes were shipped back to Toronto, however, ULDs shipped to Europe either returned to Montreal, Toronto or Vancouver All flight times varied between 1h and 9.7h (Table 6 1). ULD capacity As stated in Table 6 1, domestic fli ghts were made onboard Airbus 320 or 321 narrow body aircrafts, which hold LD3 45 ULDs (or called by their prefix

PAGE 135

135 AKH). These ULDs take the entire width of the cargo hold. The A320 can hold three in the forward compartment and four in the aft compartment. Being slightly longer, the A321 can hold five ULDs i n the front as well as four in the aft cargo hold. On the other hand, international flights were made on wide body aircrafts which carry LD3 ULDs (prefix AKE). Those ULDs occupy half of the cargo hold wid th and are loaded side by side. The A330 can hold 18 in the front and 15 in the aft compartment, whereas the Boeing 777 can carry up to 24 in the front and 20 in the aft compartment. Five tags were placed in each AKH and eight tags were installed in each A KE container according to the scheme shown in Figure 6 1. Temperature control According to Air Canada Load Control Engineering publications t here is no ventilation and no heating in any of the cargo hold s of the A320 and A321 aircraft s except for cargo door leakage when there is a pressure differential between the fuselage interior and exterior. The Airbus specification for t he A321 aircraft guarantees a minimum temperature of 2 C in flight (Air Canada, 2005a, b ) In the A330, t he forward cargo hold is e quipped with a temperature controlled heating and cooling system that is ventilated at all times. U nder normal conditions, the mean in flight temperature will be between 5 C and 25 C The aft cargo compartment is neither provided with ventilation nor heati ng system (Air Canada, 2006) In the B777, a ll cargo holds are heated Th e forward hold is equipped with an air conditioning system designed to maintain a constant target temperature and provide ventilation both on the ground or during flight. A temperatur e selector in the cockpit provides a selectable temperature control ranging from 4 C to 27 C. Th e aft hold is equipped with a basic heating system providing compartment temperature control to two set points

PAGE 136

136 corresponding to settings of LOW ( 4 C to 10 C ) an d HIGH (18 C to 24 C ) (Air Canada, 2007) Temperature sensors The temperature sensors used for monitoring were TurboTag (Sealed Air Corporation, Elmwood Park, NJ), which are high frequency (13.56MHz) RFID tags with temperature logging capacity of 702 tim e temperature data points. Tag accuracy is 0.5C throughout n ormal o perating r ange ( 25C to +35C) at 95 % confidence interval. They were programmed to read every five minutes, which allowed close to 2.5 days of monitoring time. All tags were started the morning before the first flights and stopped automatically when all 702 points were recorded. Data was downloaded at the end of the experiment. between the time the aircraft left the origin gate, and the time it sets the brakes at the destination gate, thus including taxi, takeoff and landing. Moreover, airside operations refers to everything between the time the cargo leaves the warehouse to go on the tarmac, until it comes back to another warehouse for customer pick up. Results and Discussion During Flight On a general basis, sensors recorded much lower temperatures in the bottom of ULDs than in the upp er part during flight (Figure 6 2) This observation supports one of the initial hypotheses. Moreover, t his fact can be explained by a combination of factors such as the distance from the aircraft skin, the heat coming from the passenger cabin and natural convection. The narrow body aircrafts used in this study were the Airbus 320 and 321 and are considered as short to medium range aircrafts. As explained earlier, these aircrafts do

PAGE 137

137 not have any mean s for heating their cargo holds, and consequently temperat ure distribution depends entirely on outside/surrounding conditions For narrow body aircrafts, this study shows that the duration of the flight significantly affects temperature drop (Figure 6 2). Flights under 2h (Toronto Montreal) stayed above 15 C all around (Figure 6 3), whereas flights to and from Vancouver (4 6h) dropped to 3 C at the bottom of the ULDs (Figure 6 4). Mostly, it is the temperature on the floor of the cargo hold that cools the most, but the overall temperature also drops. Cold or warm temperature does not necessarily mean good or bad. Temperature sensitive goods do not always require refrigeration. Tropical fruits, for example, should never be exposed to close to freezing temperatures, and berries, on the other hand, should be kept as c lose to 1 C as possible. For wide body aircrafts (A330 and B777 ), temperature shows a decreasing trend while in the air (Figure 6 5 and 6 6). When averaging all curves on Figure 6 5, the general slope becomes negative after only 1h. As a result, flight ti me is inversely proportional to overall temperature in the cargo compartments, at least for flights over 7h. Nonetheless, the temperature drop during flight was steeper for the narrow body (Figure 6 4) for the same length of time when comparing with the fi rst 5h of wide body 5, 6 6). This could be due to the temperature control ability of wide body aircrafts. However, the only ULD carried in the narrow body A321 (Figure 6 4) was placed at the same position on both flights (Figur e E 2), therefore results could change with data from other locations inside the aircraft. For shipments in the A330, some ULDs were carried aboard the forward or aft

PAGE 138

138 were no t. Averaging all tag temperatures carried in the forward (heated) compartment during flight compared to those transported in the aft (unheated) cargo hold (Table 6 2) leads to the conclusion that there is no significant difference in temperature profiles. This conclusion might be biased by the fact that all ULDs transported in the heated compartment were positioned at the very front of the aircraft, right next to the cargo door. Moreover, if there was no requirement for heating the cargo hold during those s no heat. Those two factors could have strongly contributed to this almost homogeneous result which is counterintuitive. Before and After Flight ULDs are brought to t he gate up to 2h before each flight, and can typically spend the same amount of time waiting on the tarmac after being unloaded from the aircraft (an ideal, no delay situation ) (T. Howard, 2010, personal communication). But as the present study showed, eve n a short period of time waiting on the tarmac can lead to very high temperatures on top of the UDLs (in summer season). Maximum recorded temperatures at the top of the ULD peaked at 45 C in less than 20 minutes from arrival at the gate (see 7h mark in Fig ure 6 7). The goods pulp temperature certainly do not warm up as fast, but sensitive shipments placed near the top of the containers could easily suffer from a major break in the cold chain. On the other hand, goods loaded in the bottom part of the contain er would take much longer to experience the temperature profile (2h before first flight until 2h after second flight) is plotted separately. Each graph also shows the container p ositions inside the aircrafts. Throughout the year, the effect of

PAGE 139

139 ground weather on ULD temperature should see much more variation than during the in flight transportation since air temperature s at high altitude s do not vary as much. Conclusion Cold chain is always hard to keep intact when goods transfer from one hand to another, especially when the transportation methods do not offer refrigeration during transit. This temperature distribution study showed that a major temperature gradient can be found wit hin the same ULD during tarmac operations as well as during flight, especially when the flight time exceeds 4h. Moreover, temperature seems to drop faster inside the cargo holds of narrow body aircrafts, but further similar studies are required to verify t hat statement. Since this study was performed in the summer, it would be both interesting and informative to measure temperature distribution and variability for similar tests performed during the winter. Furthermore, one could investigate transportation w ith longer flights (>10h) to measure temperature variability and distribution and to see if temperature in the bottom of the ULDs would reach critical freezing temperatures.

PAGE 140

140 Table 6 1. Routes, aircraft and ULD specs from Toronto (YYZ). Destination 1 Des tination 2 ULDs qty type Aircrafts to/from Flight time to / from (h) YUL YYZ 3 AKH Airbus 320 1.1 / 1.4 YVR YYZ 1 AKH Airbus 321 5.4 / 4.3 LHR LHR LHR YUL YYZ YVR 2 1 1 AKE AKE AKE Airbus 330 Airbus 330 A330/B777 7.2 / 7.5 7.2 / 8.0 7.2 / 9.7 FRA Y UL 4 AKE Boeing 777 7.5 / 7.5 Table 6 2. Temperature comparison between heated and unheated cargo holds inside an Airbus 330. Heated compartment Unheated compartment F light AKE # Temperature mean s ( C) F light AKE # T emperature mean s ( C) Top tags bo ttom tags Top tags bottom tags YYZ LHR 0 3782 19.20 12.71 YYZ LHR 0 4090 17.60 13.97 YYZ LHR 0 4969 21.26 15.58 LHR YYZ 0 4090 20.20 14.71 LHR YUL 0 4969 20.78 13.18 YYZ LHR 0 5335 17.54 14.01 LHR YUL 0 5335 22.02 16.79 Total averages 20. 41 13.82 Total averages 19.34 14.87 17.12 17.11 Figure 6 1. ULD types and their respective tag positions

PAGE 141

141 Figure 6 2. Temperatures recorded for top and bottom tags during flight (gate to gate). Data is congregated by total fligh t time and type of aircraft. The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. E rror bars above and below the box ind icate the 90th and 10th percentiles.

PAGE 142

142 Figure 6 3. Graph of averaged top and bottom tag temperatures during flight (gate to gate) for both short flights (1 2h) to and from Montreal (YUL). Figure 6 4. Graph of averaged top and bottom tag temperatures during flight (gate to gate) for both medium short flights (4 6h) to and from Vancouver (YVR).

PAGE 143

143 Figure 6 5. Graph of averaged top and bottom tag temperatures during flight (gate to gate) for all 7 8h flights to and from London (LHR) or Frankfurt (FRA). R ed pink colors are for top temperatures, and blue green colors are for bottom temperatures. Figure 6 6. Graph of averaged top and bottom tag temperatures during flight (gate to gate) for the longest flight (above 9h) between London (LHR) and Vancouver ( YVR).

PAGE 144

144 Figure 6 7. Temperature profiles of all tags for ULDs AKH 1817 to and from Montreal. First flight segment is highlighted in yellow and returning flight segment is highlighted in blue. Corresponding container positions are shown on the sketch to th e right.

PAGE 145

145 CHAPTER 7 GLOBAL TRACKING SYST EM FOR AIR CARGO SUP PLY CHAIN Introduction Previous chapters have analyzed the use of radio frequency identification technology from the view point of air cargo transportation by exploring a wide variety of factors such as different materials, frequencies and implementations. The next logical step would be to describe the advantages radio frequency identification (RFID) brings to air cargo transportation and utilize these findings to recommend important parameters o f a functional RFID prototype system to be used for air cargo tracking. Even though there are a few RFID applications in air cargo business asid e from some local solutions and smaller pilot projects for testing purposes, no major applications yet exist in this arena (Chang et al., in press) According to the Merriam Merriam Webster ). In the case of general air cargo system, the items are the g oods or freight, which can be grouped into unit load devices (ULDs) whereas the unified whole constitutes the entire distribution chain. The air cargo distribution chain starts in the cargo warehouse where the goods are accepted, after which they are loade d into ULDs, brought to the ramp, loaded onto the aircraft and flown to destination. The chain stops when the goods are picked up at the other end. When implementing an RFID tracking system in the air cargo supply chain, it can be seen as a subsystem of th e main transportation system Therefore, for the air cargo RFID tracking system, the RFID tags are the items, the ULD tags represent the group of items and the readers and the infrastructure make the unified whole.

PAGE 146

146 Goal of the system The global air cargo RFID system should be able to gather and provide valuable information (items description, weight, dimension and location, time stamps of freight movement, alarms for temperature abuse, etc) along the transportation chain in an easy and effortless manner. T he placement of RFID tags and readers should not be intrusive and should help improve operation efficiency and precision. The idea is to upgrade the quality of information without clogging or overflowing the databases with useless numbers. The objective of this chapter is to identify the applicable points of RFID technology in the air cargo handling process. The main expected benefits would include: Ability to track shipments at the item level (shipment visibility) Improve operational performance and effici ency (simplify processes, manage recoveries and decrease processing time) Improved customer experience (shipment visibility online) Minimize reliance on manual input (r educe claims and performance failure) However, for the implementation to give the best success it is necessary to evaluate carefully which technology is best in terms of frequency, type of tags, surrounding materials, etc. The findings in this dissertation will serve as a guideline when providing these recommendations. The following sections will describe the potential use case of RFID for each individual air cargo operation in greater detail. Typical Air Cargo Warehouse Operations The general process of freight handling as seen today in an air cargo warehouse is described in details in tabl es 7 1 to 7 5, while a brief overview is presented in Figure 7 1. The left columns of the tables describe the current process, whereas the right column explains how an RFID tracking system could improve the process. The actors

PAGE 147

147 taking part are: cargo agent, station attendant, cargo planner, lead agent, booking coordinator, consignee and shipper. The process ameliorations described in tables 7 1 to 7 5 can have more positive consequences than what they were originally designed for. In process #2 (Table 7 1) tagging every individual piece of a shipment with an RFID tag, instead of tagging every piece with the exact same label, allows unique identification of each piece as well as weight management. This is useful further down the road when they are loaded ont o ULDs (process #7, Table 7 2). By knowing the weight of each single item it would be possible to estimate the ULD total weight more accurately, and therefore give a better n the cargo hold. In contrast, when a shipment of 10 boxes is accepted today, only the collective weight of all the boxes is recorded and every box has the same indentifying label (airway bill number, destination, client information, etc.), apart from show be efficient for the planning agent (process #5, Table 7 2). For example, he could know ahead of time when an odd shaped package has to be planned onto a pa llet because it would simply not fit inside a container. ULD/item association is currently done manually and ineffectively by writing down which shipment is in which ULD (process #7, Table 7 2). Automatic association via RFID would minimize the risk of hum an error or hard to read messy hand writing. It could also be used to tell the station attendant if something is missing or if something should not be loaded right away. ULD tagging is not only useful for item/ULD association (process #7, Table 7 2) and UL D movement tracking in and out of the

PAGE 148

148 warehouse (processes #11 and #12, Table 7 3), but also allows faster item locating and easier recovery if they were not placed in the right area. Each year, larger cargo airlines lose 5 6% of their ULD inventory amou nting to hundreds of millions of dollars in loss due to breakdowns in their ULD tracking facilities ( Skorna and Richter 2007 ). Typical Air Cargo Ramp Operations When ULDs and tub carts are brought to the ramp prior to loading of the aircrafts, some unpr edictable and undesirable events can occur For instance, the cargo could be dropped off at the wrong gate, or c onversely, when the aircraft is being unloaded, the cargo could sit on the tarmac for a long time if the runner got the wrong message or forgot one of the ULDs at the ramp. Immediate k such mistakes and optimization of the operations. Moreover, temperature tracking of perishable cargo on the tarmac could help prioritize the movement of goods by setting off alarms when sensitive products are being exposed to extreme conditions. This study showed that even short periods of time on the tarmac could lead to high temperature elevations at the top of the ULDs (see chapter 6) for summer months. The opposite is probably true during winter, as the outside temperature is well below freezing the freight can face highly damaging environment. Either way, cargo being exposed to outdoor conditions of all kinds is very vulnerable to te mperature abuse and should be monitored to improve quality control. Ramp operations are managed by airport employees, so when the air cargo company delivers the goods to the ramp the managing of the goods is out of their control until they are unloaded at destination. Goods can be delivered to the ramp up to 2h in advance of flight schedule. If the flight is delayed, the goods can stay on the

PAGE 149

149 tarmac for long periods of time. When the weight and balance calculation is ready, the ground crew can start loadin g the cargo inside the plane according to the loading plan. The introduction of an RFID system would allow automatic time stamps and confirmation that cargo is on board (for operation efficiency and customer information update). Furthermore, in case a ULD is bumped (not flying on schedule) due to flight overweight for example, a message could be sent immediately to inform parties of this situation. Thus, using an RFID tracking system would help reduce the time that sensitive cargo may spend sitting on the t armac under diverse weather conditions and sometime for long periods of time. In addition, RFID instrumented cargo holds could permit temperature monitoring of sensitive goods. Or in a simpler application, it could notify the pilot of the recommended temp erature to set the cargo hold at during flight based on the transported goods and the required temperature range information recorded on the tags. Currently, only aircrafts with temperature control capabilities are equipped with temperature sensors in thei r cargo hold (Howard, 2010, personal communication ). Temperature monitoring would not only benefit perishables or live animals; it could also act as a back up security system against adverse situations in the cargo hold such as fires. Findings from this St udy and Recommendations for RFID Tracking System Implementation RFID implementation necessitates the incorporation of many factors and variables and the corresponding optimization based on the unique properties of the implementation environment. Even thoug h the scope of the study discussed in this dissertation has not included all the aspects required for a successful commercial RFID implementation, it still provides invaluable information on some of the major variables

PAGE 150

150 when implementing an RFID system such as choosing the right tag, the right frequency and the right ULD material. Following sections will describe these recommendations in greater detail. Passive and Active RFID Tags RFID systems can either be passive, which mean they communicate via signal ba ckscattering and rely on RF energy transferred from the reader to the tag to power the tag; or they can be active or semi active meaning the tags have their own internal power source ( typically a battery) to continuously power the tag and its RF communica tion circuitry, leading to longer read ranges than for passive tags. For the purpose of simplification, in this text, the word active will be used to represent both active and semi active RFID tags. RFID tags used for piece level identification are only be ing used once, and consequently have to be cheap. On the other hand, ULD tags can be permanently installed and reprogrammable, and therefore justifies a higher cost. From a practical point of view for air cargo, active tags are more appropriate for ULD ide ntification because of their bigger size and longer read range. For similar reasons passive tags are better suited for piece level identification due to their smaller size and affordability. Moreover, active tags can easily support sensor applications such as temperature monitoring As was observed in chapter 6, cargo encounters wide temperature variations during transit, hence the need for a temperature monitoring solution. Temperature control inside the aircraft would be made possible with the use of acti ve RFID temperature tags. RF propagation characteristics evaluated in this study (see chapter 4) would permit much better reader tag communication for active systems than passive

PAGE 151

151 systems. To achieve a RFID tag read, two communication links must be successf ul. First, the reader to tag link must not fail, and second, the tag to reader link has to be complete. In the case of passive systems, the success of the communication is often limited by the reader to tag link ( Nikitin and Rao, 2006; Dobkin, 2008). The r eader has a specific sensitivity in the order of 65dBm to 120dBm, whereas the passive tags have sensitivities of around 1 2 dBm (Nikitin et al., 2009). Therefore, when the reader sends a signal into its surrounding and the signal is attenuated with dist ance and interfering objects if the tags receive enough energy to respond, their signal will most likely be received with enough energy at the reader end as well. In the case of active RFID systems, the tags do not need to collect a minimum amount of ene rgy from the reader to broadcast. Having their own power source allows them to transmit their information to other words, it is the sensitivity of the reader that will determine the success of the communication, and therefore permit a much longer read range than for passive systems. The signal levels observed in chapter 4 were from 16 to 27dBm for 433MHz, 6 to 17dBm for 915MHz, and 18 to 27dBm for 2.45GHz (Appendi x D). If a typical passive RFID tag has a sensitivity or threshold of around 12dBm (Nikitin et al., 2009), even 915MHz would not offer enough coverage to read the tags anywhere in the cargo hold when using this test set up (one reader antenna). Moreover, this test was performed inside an empty cargo compartment, which means the signal strengths would most likely be further reduced in the case of a fully loaded cargo hold. Further testing is required, but so far active RFID systems are thought to be a more feasible solution.

PAGE 152

152 ULD Materials Air cargo containers or ULDs can be made of different materials. Older ULDs were all made of aluminum, whereas newer ones are made of Kevlar composite walls on an aluminum frame. Old containers are being replaced by this n ew style because of their much lighter weight ( Howard, 2010, personal communication ; Nordisk, 2010). This change is favorable to the findings of this study (see chapter 5), which state that Kevlar is highly RF lucent, whereas aluminum is RF opaque. In oth er words, Kevlar lets RF waves go through when aluminum totally reflects them. This result suggest that most RFID tags applied to the surface of Kevlar UDLs would presumably lead to better readability than those applied on aluminum ULDs. Moreover, in the case of temperature monitoring, using Kevlar ULD holds a strong advantage over aluminum since their content information could be read directly through the walls. Frequency Warehouse The warehouse environment is susceptible to many external noises, such as 2 way radios, wireless networks, cell phones and automatic door entry systems, which could interfere with the RFID communication links as described in chapter 3. Based on this study, the frequencies of choice are 433MHz and 915MHz because of lower inte rference levels around these bands. Using 915MHz would give the flexibility of using active and/or passive RFID systems because of the higher power output allowance by FCC regulations in this band (FCC, 2008); whereas 433MHz is only suitable for active RFI D systems. Using both active and passive systems could be a plausible solution as well given the use case requirements. As mentioned earlier, item tags have to be cheap because they are generally not being reused in the system, whereas container tags can b e reprogrammable and permanently installed on the unit.

PAGE 153

153 Taking into consideration the interference levels and the flexibility of the systems, 915MHz seems to be the best option for passive RFID, but both 433MHz and 915MHz could serve for the active. 2.4GHz displays a more significant interference due to wireless and GSM networks and thus not recommended for warehouse implementation. Aircraft The frequency of choice for cargo hold identification was shown to be 915MHz (see Chapter 4) because of an allowed m aximum output power higher than at 433MHz (FCC, 2008) as well as a lower attenuation compared with 2.45GHz. As stated earlier, the higher signal level makes passive RFID systems a possible solution. On the other hand, when considering active systems, 433MH z might be a good alternative due to lower attenuation However, before RFID implementation can take place inside an aircraft, many studies will have to be conducted to prove that RF signal would not be significantly interfering with other aircraft radio s ystems as per the Federal Aviation Administration document AC20 International c ompatibility ISM bands around 433MHz and 2.45GHz are available internationally. On the other hand, as much as 915MHz seems to be a good solution for many different applications it is only allowed for use in the Americas e rest of the world has different regulations for using similar frequencies, which can fall between 860 960MHz (standard ISO/IEC 18000 6). Therefore, a global air cargo tracking system would have to account for all those frequencies to be implementable eve rywhere. There exists tags that can function anywhere within that range, but the different regulations can imply different maximum output powers, different bandwidths, etc. Consequently, a

PAGE 154

154 system that works in the United States would not necessarily work i n an id entical manner elsewhere. Such limitations would need to be taken into consideration in detail before a global system can be designed with assurance. Conclusion This chapter draws upon the results and conclusions of previous chapters to list the adv antages and important parameters of a functioning RFID system from the view and an RFID enabled version is presented to show how greatly RFID would improve the visib ility throughout the entire supply chain. In addition, the findings of previous chapters such as the interaction of air container materials with RFID, the advantages of different RFID frequencies in different situations, etc. are utilized to recommend par ameters (such as the type of tag or the frequency band) based on the use case scenario. It is important to note that a fully functional implementation of RFID in air cargo supply chain would require full collaboration of many different parties involving pr ivate companies and government institutions. However, when the advantages presented in this chapter are taken into account, it is trivial to see such efforts would be beneficial for the entire air transportation industry.

PAGE 155

155 Table 7 1. Current processes as well as proposed RFID solutions for the air cargo supply chain (cargo acceptance part). Cargo acceptance # Improvement with RFID system 1 Cargo agent determines origin and destination of requested service, determines compatibility of sh ipment with aircraft and station, and if acceptable, confirms booking with customer. 2 Station attendant receives freight from customer and verifies size, weight and number of pieces at acceptance dock. Information is hand written on a piece of paper. Pr oduct information is associated with the RFID tag ID which is applied to every piece of a shipment. Each piece has its own weight, dimension and special requirements (temperature, dangerous goods, live animal, priority, etc.) info associated with its RFID tag. Option: Weighing, dimensioning, tag programming and application could all be achieved automatically via a conveyor belt system. 3 Cargo agent verifies documents and accepts cargo if acceptable. He creates an airway bill (AWB), prints and places labe ls in a tray for station attendant. Tag ID and AWB num ber are associated by reading the tag with a handheld RFID reader. 4 Station attendant attaches labels to shipment and delivers it to build up area. Labels have already been applied to shipment.

PAGE 156

156 Table 7 2. Current processes as well as proposed RFID solutions for the air cargo supply chain (cargo build up part). Cargo build up # Improvement with RFID system 5 Cargo planner creates build up plan. 6 Booking coordinator pulls bui ld up plan. 7 Station attendant loads shipment as per build up plan and creates a planning load assembly (PLA) concurrently. Container ULD: Each ULD has its own RFID tag, which is scanned simultaneously as the items are being loaded inside This creates automatic a ssociation of the item level pieces and ULDs. This could be achieved with a wearable or handheld RFID reader. Pallet ULD: Build up areas for pallet are predetermined by roller system on the floor. Therefore, item association could be done with a fixed RFID reader on the ceiling above the pallet build up pit. Note: All ULD tags include their tare weight information so that shipment association gives a precise estimate of the ULD total weight after build up. 8 Booking coordinator checks for add itional shipments, adds to and finalizes build up plan. 9 Station attendant prepares final loads and PLAs and sends them to the planners 2h prior to the flight schedule. The information is sent to the planner automatically from the database as the ULDs a re created. 10 Planner enters info r mation in the database This step is done automatically since all cargo information was updated in the database during build up.

PAGE 157

157 Table 7 3. Current processes as well as proposed RFID solution s for the air cargo supply chain (cargo to/from the ramp section). Cargo to/from the ramp # Improvement with RFID system 11 S tation attendant stages and runs freight to aircraft 2h prior to flight schedule All ULD IDs are automatically read when crossing the warehouse export doors (portal RFID reader). Database is updated of the goods departure. (Cargo delivered at the ramp is now in the hands of airport employees) 12 S tation attendant checks teletype for any special commodities me ssages (for inbound freight), retrieves time sensitive goods first and delivers to warehouse and informs Lead agent. Station attendant receives an alarm when time sensitive goods are ready for pick up. An additional message tells him the required temperat ure of the item for proper storage location in warehouse. All ULDs are automatically read at warehouse inbound door. Database is automatically updated of the goods arrival. 13 S tation attendant retrieves and delivers non time sensitive goods to import sid e of warehouse. All ULDs were automatically read at warehouse inbound door. Database is automatically updated of the goods arrival. 14 C argo agent (In flight coordinator) begins database check in, prints inbound manifest and verifies with physical goods. Already done by automatic reading of the goods through inbound doorway entrance.

PAGE 158

15 8 Table 7 4. Current processes as well as proposed RFID solutions for the air cargo supply chain (cargo break down and storage section). Cargo break down and storage # To Improvement with RFID system 15 S tation attendant sends co py of inbound manifest to lead agent retrieves non shipper loaded units from storage, breaks down and sorts by AWB, moves to appropriate location, scans notes location on AWB and r et urns completed manifest to cargo agent. During break down, every piece is disassociated from the ULD, the same way it was associated previously. Pieces are read and associated with their storage location 16 S tation attendant moves shipper loaded units ( S LUDs ) and connecting shipments to proper location. ULD t ags should be read and associated with their waiting location 17 C argo agent checks in SLUDs in database. Automatic with scanning in previous step. 18 C argo agent check in goods not previously sca nned determines if pieces or entire shipments are missing, performs missing cargo transaction and completes check in process Missing cargo should be triggered automatically since all goods were read at the inbound door. 19 C argo agent identifies perisha ble, Live or hold for pick up contacts customer electronically or by phone, records conversation, completes paperwork and waits for customer to contact back If no contact within 14 days warning is mailed to consignee. If no contact within 30 days, final warning is mailed to consignee and shipment is destroyed if destination is domestic or reported if not domestic. Alarm is sent to cargo agent when time sensitive shipments have entered the warehouse. Customer receive an automatic email when their goods h ave arrived at the warehouse and are ready for pick up. 20 C argo agent determines if customer are perishable, live or priority, cargo agent informs customer of 24h p ick up window, otherwi se informs customer of 48h pic k arranges for ground transportation as per customer priority. Automatic from database through email.

PAGE 159

159 Table 7 5. Current processes as well as proposed RFID solutions for the air cargo supply chain (cargo delivery part). Cargo del ivery # Improvement with RFID system 21 C argo agent delivers document s to broker or consignee if delivery is requested and goods are not domestic, for Customs clearance. Once cleared, cargo agent checks AWB for Customs stamp. 22 C argo agent obtains consignee signature on AWB and collects outstanding charges for goods picked up, co mpletes delivery process in database and identifies warehouse consignee. 23 S tation attend ant receives AWB from stamp and retrieves shipment. Before goods are released, an RFID read is required to update the database. 24 S tation attendant ensures that consignee inspects shipment for damage before delivery a nd releases to if not damaged. If damaged, completes bad order report and attach to AWB. Final database update, case closed. Figure 7 1. Overview of the air cargo operations where the circled steps represent suggested RFID reading points.

PAGE 160

160 CHAPTER 8 GENERAL CONCLUSION The goal of this dissertation was to explore the possibility of using radio frequency identification ( RFID ) to improve air cargo operations in terms of efficiency, safety and monitoring. This study showed interference levels at three u ltra high frequencies (UHF) recorded in two air cargo warehouses. The interference levels from highest to lowest were at 2.45GHz, 433MHz and 915MHz respectively. There are ways to filter out interferences when designing an RFID system, high end readers have that feature. However, the best way to avoid possible disturbing noises is surely to install the RFID system in an interference free environment According to the results of these tests, which were performed inside two warehouses whic h may or may not give a realistic representation of all air cargo warehouses in the world, implementation of RFID systems at 915MHz in North America would bring the best results, interference wise. This study also demonstrate d that frequencies have a majo r influence on signal propagation, especially inside a metal environment. Lower frequencies suffer less attenuation over distance, but have higher variation within the cargo hold. It was also demonstrated that antenna polarization can have a significant ef fect on signal propagation in some cases, and therefore should not be omitted when designing an RFID system for air cargo transportation. Moreover, FCC regulations restricts output powers at 433MHz more than at 915MHz and 2.45GHz, leading to the conclusion that more RF energy would be avai lable in the cargo hold for reader /tag communication at 915MHz than at the other frequencies tested. Moreover, the study showed that the relationship between signal strength and tag reads is an important tool to take into account when implementing RFID systems

PAGE 161

161 This dissertation verified the effects of five commonly used air cargo container wall materials on RF propagation at the same three different frequencies. Three different tests were utilized to analyze the characteri stics of RF propagation for each material and the results from all experiments showed a very strong effect of aluminum on RF transmission and minimal interaction for all other sample materials as expected. These finding suggest that the use of non metallic containers for air transportation of perishable products should make real time temperature monitoring possible by allowing RF waves to transmit through the wall surface effortlessly. This study demonstrated that a major temperature gradient can be found w ithin the same ULD during ground operations as well as during flight, especially when the flight time exceeds 4h. Therefore, it is suggested that temperature sensitive shipments should be placed accordingly inside the ULD. This test was performed in the su mmer, for that reason, the increase of temperature during ground operations should be considered variable. However, the temperature distribution observed during flight should be consistent through the year since temperatures at high altitude do not vary wi dely. This work presented a an RFID enabled version to show how greatly RFID would improve the visibility throughout the entire supply chain. It is important to note that a fully functional implemen tation of RFID in air cargo supply chain would require full collaboration of many different parties involving private companies and government institutions. However, when th e advantages presented in this dissertation are taken into account, it is trivial t o see such efforts would be beneficial for the entire air transportation industry.

PAGE 162

162 APPENDIX A DC 10 CARGO HOLD AND CARGO DOOR SPECS Figure A 1. Standard cargo compartment and containers for model DC 10 series 10, 10CF, 30, 30CF, 40 and 40CF (Boeing, 2010c)

PAGE 163

163 Figure A 2. Forward cargo loading door, model DC 10 series 10, 10CF, 30, 30CF, 40 and 40CF (Boeing, 2010c)

PAGE 164

164 APPENDIX B R ADIO F REQUENCY ATTENUATION SURFACE PLOTS All plot s are following the color coded spectrum where red is high attenuation, o r 70 dBm (weak signal) and p urple is low attenuation, or 30 dBm (strong signal) as indicated in the example graph below. Slices are numbered from 1 to 12 which represent the distance from the front of the cargo hold in meters. Surface plots are shown as if y ou were standing at the back of the aircraft, looking forward. Therefore, the right side of the plot is the starboard side of the vessel and the left is port, as also indicated in the example plot below. Figure B 1. Attenuation s urface plot ex ample one slice of data (dBm) Port Center Starboard High Middl e Low Tripod sideways position inside the cargo hold Tripod height

PAGE 165

165 Figure B 2. Attenuation s urface plot for 433MHz, top end a ntenna position and c ircular p olarization

PAGE 166

166 Figure B 3. Attenuation s urface plot for 433MHz, center ceiling a ntenna position and circul ar p olarization

PAGE 167

167 Figure B 4. Attenuation s urface plot for 915MHz, top end a ntenna position and c ircular p olarization

PAGE 168

168 Figure B 5. Attenuation s urface plot for 915MHz, top end a ntenna position and line ar p olarization

PAGE 169

169 Figure B 6. Attenuation s urface plot for 915MHz, center cei ling a ntenna position and circul ar p olarization

PAGE 170

170 Figure B 7. Attenuation s urface plot for 2.45GHz top end a ntenna position and c ircular p olarization

PAGE 171

171 Figure B 8. Attenuation s urface plot for 2.45GHz top end a ntenna position and line ar p olarization

PAGE 172

172 Figure B 9. Attenuation s urface plot for 2.45GHz center ceiling a ntenna position and circul ar p olarization

PAGE 173

173 APPENDIX C STATISTICAL ANALYSIS RESULTS FOR DC 10 R ADIO F REQUENCY PROPAGATION All statistical analyses were computed using SAS 9.1 (SAS Institut e Inc., Cary NC) and significance was accepted Table C 1. Effects of frequency, antenna location and antenna polarization on attenuation levels of the complete dataset. E ffects mean F value p value Frequency 433 MHz 41.69 1258.26 < 0.0001 915 MHz 47.64 2.45GHz 59.34 Location Top End 51.39 15.19 < 0.0001 Ceiling 49.12 Polarization Circular 49.23 68.81 < 0.0001 Linear 54.47

PAGE 174

174 Table C 2. Effect of width on attenuation levels for each frequency, antenna location a nd polarization. Constants Effects mean F value p value Frequency Location Polarization Width 433 Top End Circular Port 42.63 2.06 0.1328 Center 42.03 Starboard 40.45 Top End Linear Port N/A N/A N/A Center N/A Starboard N/A Ceiling Circular Port 42.51 0.84 0.4343 Center 41.22 Starboard 41.30 915 Top End Circular Port 48.36 0.46 0.6331 Center 47.57 Starboard 48.06 Top End Linear Port 48.85 0.62 0.5417 Center 47.38 Starboard 47.84 Ceili ng Circular Port 46.94 0.39 0.6760 Center 46.55 Starboard 47.26 2450 Top End Circular Port 58.77 0.7 0.4992 Center 58.32 Starboard 57.87 Top End Linear Port 62.60 12.5 < 0.0001 Center 58.39 Starboard 61.75 Ceiling C ircular Port 58.11 3.04 0.0522 Center 58.47 Starboard 59.77

PAGE 175

175 Table C 3. Effect of height on attenuation levels for each frequency, antenna location and polarization. Constants Effects mean F value p value Frequency Location Polarization Hei ght 433 Top End Circular High 41.51 0.25 0.7770 Middle 41.43 Low 42.16 Top End Linear High N/A N/A N/A Middle N/A Low N/A Ceiling Circular High 41.11 0.46 0.6353 Middle 41.74 Low 42.18 915 Top End Circular High 48.43 2.16 0.1210 Middle 47.01 Low 48.54 Top End Linear High 52.66 27.13 < 0.0001 Middle 46.37 Low 45.03 Ceiling Circular High 46.86 0.01 0.9936 Middle 46.94 Low 46.94 2450 Top End Circular High 58.25 0.94 0.3943 Middle 58.87 Low 57.85 Top End Linear High 61.91 1.61 0.2054 Middle 60.28 Low 60.54 Ceiling Circular High 59.16 0.8 0.4524 Middle 58.92 Low 58.28

PAGE 176

176 Table C 4. Effect of depth on attenuation levels for each frequency, antenna location and polarization. Constants Effects mean F value p value Frequency Location Polarization Depth (dBm) 433 MHz Top End Circular 1 36.61 6.28 < 0.0001 2 38.97 3 37.73 4 39.65 5 39.15 6 42.32 7 43.51 8 43.30 9 44.41 10 43.47 11 44.21 12 47.11 Top End Linear 1 N/A N/A N/A 2 N/A 3 N/A 4 N/A 5 N/A 6 N/A 7 N/A 8 N/A 9 N/A 10 N/A 11 N/A 12 N/A Ceiling Circular 1 46.96 3.61 0.0003 2 44.10 3 40.47 4 39.97 5 40.82 6 36.92 7 39.03 8 40.08 9 42.96 10 43.24 11 42.58 12 42.93

PAGE 177

177 Table C 4. Continued Constants Effect Mean F value p value Frequency Locati on Polarization Depth (dBm) 915 MHz Top End Circular 1 43.63 8.33 < 0.0001 2 45.48 3 45.99 4 46.45 5 47.21 6 47.67 7 47.94 8 48.72 9 48.28 10 50.95 11 52.67 12 50.94 Top End Linear 1 42. 22 4.88 < 0.0001 2 44.54 3 44.94 4 45.02 5 45.40 6 47.32 7 49.07 8 49.64 9 52.04 10 51.61 11 52.48 12 51.94 Ceiling Circular 1 50.81 11.82 < 0.0001 2 49.51 3 46.75 4 47.27 5 44.68 6 43.25 7 42.37 8 45.15 9 46.86 10 47.35 11 43.59 12 50.32

PAGE 178

178 Table C 4. Continued Constants Effect Mean F value p value Frequency Location Polarization Depth (dBm) 2.45GHz Top End Circular 1 5 4.57 5.83 < 0.0001 2 55.44 3 56.77 4 58.37 5 57.55 6 57.78 7 58.29 8 59.38 9 59.09 10 59.96 11 60.69 12 61.90 Top End Linear 1 60.05 0.53 0.8819 2 60.70 3 60.81 4 59.82 5 60.65 6 61.07 7 59.12 8 61.41 9 60.61 10 61.57 11 62.92 12 62.17 Ceiling Circular 1 59.99 9.95 < 0.0001 2 59.55 3 58.61 4 57.56 5 56.31 6 54.82 7 55.94 8 58.40 9 59.19 10 61.16 11 61.77 12 62.08

PAGE 179

179 APPENDIX D R ADIO F REQUENCY SIGNAL STRENGTH PROP AGATION SURFACE PLOT S All plots are following the color coded spectrum where purple is high signal strength, or 0dBm and red is low signal strength, or 40dBm as indicated in the example graph below. Slices are numbered from 1 to 12 which represent the distance from the front of the cargo hold in meters. Surface plots are shown as if you were standing at the back of the aircraft, looking forward. There fore, the right side of the plot is the starboard side of the vessel and the left is port, as also indicated in the example plot below. Figure D 1. Signal strength surface plot example one slice of data (dBm) Port Center Starboard High Middle Low Tripod sideways position inside the cargo hold Tripod height

PAGE 180

180 Figure D 2. Signal strength s urface plot for 433MHz, top end antenna position and circular polarization.

PAGE 181

181 Figure D 3. Signal strength surface plot for 433MHz, center ceiling antenna position and circular polarization.

PAGE 182

182 Figure D 4. Signal strength surface plot for 915MHz, top end ant enna position and circular polarization.

PAGE 183

183 Figure D 5. Signal strength surface plot for 915MHz, top end antenna position and linear polarization.

PAGE 184

184 Figure D 6. Signal strength surface plot for 915MHz, center ceiling antenna position and circular polarizati on.

PAGE 185

185 Figure D 7. Signal strength surface plot for 2.45GHz top end antenna position and circular polarization.

PAGE 186

186 Figure D 8. Signal strength surface plot for 2.45GHz top end antenna position and linear polarization.

PAGE 187

187 Figure D 9. Signal strength surface plot for 2.45GHz center ceiling antenna position and circular polarization.

PAGE 188

188 APPENDIX E UNIT LOAD DEVICES, TEMPERATURE GRAPHS AND POSITIONS Figure E 1. Temperature profiles of all tags for ULDs (A) AKH 2084 and (B) AKH 9778 to and from Montreal (YUL ). First flight segments are highlighted in yellow and returning flight segments are highlighted in blue. Corresponding container positions are shown on the A320 sketch to the right.

PAGE 189

189 Figure E 2. Temperature profiles of all tags for ULDs AKH 1987 to and from Vancouver (YVR). First flight segment is highlighted in yellow and returning flight segment is highlighted in blue. Container position was the same for both flight and is consequently shown in green on the A321 sketch to the right.

PAGE 190

190 Figure E 3. T emperature profiles of all tags for ULDs (A) AKE 03782, (B) AKE 04090, (C) AKE 04969, and (D) AKE 05335 to and from London (LHR). First flight segments are highlighted in yellow and returning flight segments are highlighted in blue. Corresponding container positions are shown on the A330 sketch to the right.

PAGE 191

191 Figure E 3. Continued

PAGE 192

192 Figure E 4. Temperature profiles of all tags for ULDs (A) AKE 03748, (B) AKE 04632, (C) AKE 05168, and (D) AKE 05255 to and from Frankfurt (FRA). First flight segments a re highlighted in yellow and returning flight segments are highlighted in blue. Corresponding container positions are shown on the B777 sketch to the right.

PAGE 193

193 Figure E 4. Continued.

PAGE 194

194 LIST OF REFERENCES Abad, E., F. Palacio, M. Nuin, A. Zrate D.E. Gon z les A. Juarros, J. M. Gmez and S. Marco. 2009. RFID smart tag for traceability and cold chain monitoring of foods: Demonstration in an intercontinental fresh fish logistic chain. Journal of Food Engineering 93(4): 394 399. Air Canada. 2007. Publication 167 B777 Load planning. 5. Air Canada. 2006. Publication 163 A330 Load planning. 5. Air Canada. 2005. Publication 169 A320 Load planning. 6. Air Canada. 2005. Publication 172 A321 Load planning. 5. Airbus. 2009. Airbus continues to pla y major role in composites revolution. Airbus 40Y of Innovation. Available at: http://www.airbus.com/en/corporate/innovation/40years innovation/40 years innovation items/?tx_ttnews%5Btt_news%5D=233&tx_ttnews%5BbackPid%5D=782&cHash= e0bce8b6c7 Accessed 05 /23 2010. Altunbas, A. E. 2010. Wearable RFID reader and compact fractal patch antenna design. Gainesville, Fla.: University of Florida. Anonym. 2007. Composite materials in aviation. The Airplanes Channel. Available at: http://www.airplanes.com/blog/com posite materials in aviation/ Accessed 04/09 2010 Anritsu. 2009. Fundamentals of interference in wireless networks. 11410 005361 16. Anritsu. 2008. Guide to Spectrum Analyzers. Ashford, N., C. A. Moore and H. P. M. Stanton. 1984. Airport operations. N ew York: Wiley. Bilstein, R. E 1994. Flight in America : from the Wrights to the astronauts. Rev. ed. Baltimore: Johns Hopkins University Press. Blaunstein, N. and C. G. Christodoulou. 2007. Radio propagation and adaptive antennas for wireless communicat ion links: terrestrial, atmospheric and ionospheric. Hoboken, N.J.: Wiley Interscience. Boeing. 2010 a Boeing 787 dreamliner will provide new solutions for airlines, passengers. Boeing. Available at: http://www.boeing.com/commercial/787family/background.h tml Accessed 04/09 2010.

PAGE 195

195 Boeing. 2010 b Commercial airplanes. Boeing. Available at: http://www.boeing.com/commercial/safety/index.html Accessed 04/09 2010. Boeing. 2010 c Commercial airplanes DC 10 family. Boeing. Available at: http://www.boeing.com/ commercial/dc 10/index.html Accessed 04/25 2010. Boeing. 2008. World air cargo forecast 2008 2009. Accessed 09/02 2009. Bollen, A. F., D. W. Brash and B. L. Bycroft. 1998. Air freight coolchain improvements using insulation and supplemental cooling. App lied Engineering in Agriculture 14(1): 49 53. Catto Smith, C. 2006. Thermal protection of perishable air cargo during ground handling. In International symposium on fresh produce supply chain management Cerino, A. and W. P. Walsh. 2000. Research and appl ication of radio frequency identification (RFID) technology to enhance aviation security. In Proceedings of the IEEE 2000 National Aerospace and Electronics Conference. NAECON 2000 127 135. Cerino, A., G. L. Fuller and R. W. Devereux. 1997. EMI from sour ces in the cargo bay. In Digital Avionics Systems Conference, 1997. 16th DASC., AIAA/IEEE, 4.1 9 16 vol.1. Chang, Y. S., C. H. Oh, Y. S. Whang, J. J. Lee, J. A. Kwon, M. S. Kang, J. S. Park and Y. P. Ung. 2006. Development of RFID enabled aircraft mainten ance system. In 2006 IEEE International Conference on Industrial In formatics, 224 229. Chang, Y. S., M. G. Son and C. H. Oh. Design and implementation of RFID based air cargo monitoring system. Advanced Engineering Informatics. In Press, Corrected Proof. Chen, S. C. Q. and V. Thomas. 2001. Optimization of inductive RFID technology. In Proceedings of the 2001 IEEE International Symposium on Electronics and the Environment, 82 87. Cheng, D. K. 1993. Fundamentals of engineering electromagnetics. Prentice Hal l. Chow, H. K. H., K. L. Choy, W. B. Lee and K. C. Lau. 2006. Design of a RFID case based resource management system for warehouse operations. Expert Systems with Applications 30(4): 561 576. Clampitt, H. G. 2006. RFID certification textbook. 2nd ed. Hou ston, TX: PWD Group, Inc.

PAGE 196

196 Day, D. A. 2009. Composites and advanced materials. U.S. Centennial of flight commission. Available at: http://www.centennialofflight.gov/essay/Evolution_of_Technology/composites/Tech 40.htm Accessed 01/26 2010. Devereux, R. W., G. L. Fuller and R. Schillinger. 1997. Further analysis of the CV 580 and B 727 aircraft RF coupling experiment data. In Digital Avionics Systems Conference, 1997. 16th DASC., AIAA/IEEE, 4.1 1 8 vol.1. Devereux, R. W., G. L. Fuller and R. Schillinger. 19 97. Electromagnetic susceptibility of installed avionics. In Digital Avionics Systems Conference, 1997. 16th DASC., AIAA/IEEE, 4.1 17 24 vol.1. Dobkin, D. M. and S. M. Weigand. 2005. Environmental effects on RFID tag antennas. Microwave Symposium Digest, 2005 IEEE MTT S International Dobkin, D. M. 2008. The RF in RFID : passive UHF RFID in practice. Amsterdam ; Boston: Elsevier / Newnes. Domdouzis, K., B. Kumar and C. Anumba. 2007. Radio Frequency Identification (RFID) applications: A brief introduction. Advanced Engineering Informatics 21(4): 350 355. Elias, B. 2007. CRS report for congress Air cargo security. RL32022. Ely, J. J., T. X. Nguyen, S. V. koppen, M. T. Salud and J. H. Beggs. 2003. Wireless phone threat assessment and new wireless technolog y concerns for aircraft navigation radios. NASA/TP 2003 212446. mond J. P. 2008. The cold chain. In RFID Technology and Applications, 144. ed. S. B. MilesSarma, S. E. and Williams, J. R., New York, NY, USA: Cambridge University press. mond J. P. 2007 Quantifying RFID's cold chain benefits. In RFI D Journal LIVE! 2007 Conference mond J. P., F. Mercier and M. C. N. Nunes. 1999. In flight temperature conditions in the holds of a widebody aircraft. In Proceedings of the 20th InternationalCongress of R efrigeration, Paper No. 281. ETSI. 2007. Electromagnetic compatibility and Radio spectrum Matters (ERM); Technical characteristics of RFID in the UHF Band; System Reference Document for Radio Frequency Identification (RFID) equipment; Part 1: RFID equipme nt operating in the range from 865 MHz to 868 MHz. ETSI TR 102 649 1. FAA. 2008. Airworthiness approval and operational allowance of RFID systems. AC20 1621.

PAGE 197

197 FAA. 2006. FAA aerospace forecast fiscal years 2006 2017. 1 90. FCC. 2008. Part 15 Radio freq uency devices. 1 157. Finkenzeller, K. 2003. RFID handbook: fundamentals and applications in contactless smart cards and identification. New York, NY, USA: John Wiley & Sons, Inc. Flightglobal. 2009. Aircraft profile: Airbus A320. Available at: http://ww w.flightglobal.com/articles/2008/01/01/221129/airbus a320 aircraft profile.html Accessed 09/02 2009. Forster, P. W. and A. C. Regan. 2001. Electronic integration in the air cargo industry: an information processing model of on time performance. Transport ation Journal 40(4): 46 61. Foster, P. R. and R. A. Burberry. 1999. Antenna problems in RFID systems. IEE Colloquium on RFID Technology 3/1 3/5. Friis, H. T. 1946. A note on a simple transmission formula. Proceedings of the I.R.E. 34(5): 254. Fuschini, F., C. Piersanti, F. Paolazzi and G. Falciasecca. 2008. Analytical approach to the backscattering from UHF RFID transponder. Antennas and Wireless Propagation Letters, IEEE 733 35. Griffin, J. D., G. D. Durgin, A. Haldi and B. Kippelen. 2006. RF tag anten na performance on various materials using radio link budgets. Antennas and Wireless Propagation Letters, IEEE 5(1): 247 250. Hallowell, S. F. and P. Z. Jankowski. 2005. Transportation security technologies research and development. In IEEE Military Commun ications Conference, 2005. MILCOM 2005. 1753 1756 Vol. 3. Henderson, T. 2010. Static electricity Lesson 3. The Physics Classroom. Available at: http://www.physicsclassroom.com/class/estatics/u8l3c.cfm Accessed 07/03 2010. IATA. 2002. ULD technical man ual, 17th edition. Montreal, Canada: International Air Transport Association. Impinj. 2007. UHF Gen 2 RFID readers Five factors for success. Intelleflex. 2010. The Intelleflex XC3 technology platform: An implementation of the new ISO Class 3 standard. Jacinto, C., M. Romano, A. Montes, P. Sousa and M. S. Nunes. 2009. RFID tuning methodology applied on airport baggage tracking. In IEEE Conference on Emerging Technologies & Factory Automation, 2009. ETFA 2009. 1 4.

PAGE 198

198 Jedermann, R. and W. Lang. 2007. Semi passive RFID and beyond: steps towards automated quality tracing in the food chain. Int. J. Radio Frequency Identification 1(3): 247 259. Jedermann, R., J. P. mond and W. Lang. 2007. Shelf life prediction by intelligent RFID technical limits of model acc uracy. In International Conference on Dynamics in Logistics Jedermann, R., L. Ruiz Garcia and W. Lang. 2009. Spatial temperature profiling by semi passive RFID loggers for perishable food transportation. Computers and Electronics in Agriculture 65(2): 14 5 154. Keskilammi, M., L. Sydnheimo and M. Kivikoski. 2003. Radio f requency t echnology for a utomated m anufacturing and l ogistics c ontrol. Part 1: Passive RFID s ystems and the e ffects of a ntenna p arameters on o perational d istance. The International Journa l of Advanced Manufacturing Technology 21(10): 769 774. Ketzenberg, M. E. and J. M. Bloemhof Ruwaard. 2009. The value of RFID technology enabled information to manage perishables. ERIM Report Series Reference No. ERS 2009 020 LIS. Khan, M. A., M. Sharma and P. R. Brahmanandha. 2009. A survey of RFID tags. International Journal of Recent Trends in Engineering 1(4) Kim, D. Y., H. G. Yoon, B. J. Jang and J. G. Yook. 2008. Interference analysis of UHF RFID systems. Progress i n Electromagnetics Research B 41 15 126. Kumar, R., S. K. Kaura, A. K. Sharma, D. P. Chhachhia and A. K. Aggarwal. 2007. Knife edge diffraction pattern as an interference phenomenon: An experimental reality. Optics & Laser Technology 39(2): 256 261. LaBerge, E. F. C. 2007. An analysis o f the effects of RFID tags on narrowband navigation and communication receivers. Lahiri, S. 2006. RFID sourcebook. Upper Saddle River, NJ, USA: IBM Press. Landt, J. 2005. The history of RFID. Potentials, IEEE 24(4): 8 11. Laniel, M., J. P. mond and A. E. Altunbas. 2009. RFID behavior study in enclosed marine container for real time temperature tracking. Sensing and Instrumentation for Food Quality and Safety 3(1): 34 40. Liang, B., A. Hu and Z. Qin. 2006. Trends and brief comments on anti collision tec hniques in radio frequency identification system. In 2006 6th International Conference on ITS Telecommunications Proceedings 241 245. Littell, R. C., G. A. Milliken, W. W. Stroup, R. D. Wolfinger and O. Schabenberber. 2006. SAS for mixed models. 2nd ed. Cary, NC, USA: SAS Publishing.

PAGE 199

199 Longhurst, R. S. 1967. Geometrical and physical optics. 2nd ed. London: Longman. Mac Carthy, U. M. 2009. Assesment of UHF RFID as an automatic ID data carrier for item level food supply chain traceability. PhD diss. Dublin: University College Dublin. Mayer, L. W., M. Wrulich and S. Caban. 2006. Measurements and channel modeling for short range indoor UHF applications. In First European Conference on Antennas and Propagation EuCAP 2006. 1 5. McCarthy, T. M. 2003. Conductin g a market opportunity analysis for air cargo operations. Transportation Journal : May 18, 2010. McCoy, T., R. J. Bullock and P. V. Brennan. 2005. RFID for airport security and efficiency. In The IEE Seminar on Signal Processing Solutions for Homeland Secu rity, 2005. (Ref. No. 2005/11108), 9 pp. Merriam Webster, I. Merriam Webster's open dictionary [electronic resource]. Springfield, MA: Merriam Webster. M eyers, M., M. Brown, S. Patadia and S. Dua. 2007. Mike Meyers' comptia RFID+ certification passport. New York: McGraw Hill. Mo, L. and H. J. Zhang. 2007. RFID antenna near the surface of metal. In 2007 International Symposium on Microwave, Antenna, Propagation and EMC Technolog ies for Wireless Communications, 803 806. Moir, I. and A. G. Seabridge. 2001. Aircraft systems: mechanical, electrical, and avionics subsystems integration. Reston, VA; Bury St. Edmunds, U.K.: American Institute of Aeronautics and Astronautics; Professional Engineering Pub. NFPG 2008. United Kingdom frequency allocation table Ng uyen, T. X. and J. J. Mielnik. 2008. Radio frequency compatibility of an RFID tag on glideslope navigation receivers. In IEEE International Symposium on Electromagneti c Compatibility. EMC 2008 1 6. Nguyen, T. X., J. J. Ely, S. V. Koppen and M. S. Fersch. 2008. RFID transponders' RF emissions in aircraft communication and navigation radio bands. In 2008 International Symposium on Electromagnetic Compatibility EMC Europe, 1 6. Nguyen, T. X., S. V. Koppen, L. J. Smit, R. A. Williams and M. T. Salud. 2005. Wireless phone threat assessment for aircraft communication and navigation radios. In 2005 International Symposium on Ele ctromagnetic Compatibility. EMC 2005 135 140 Vol. 1.

PAGE 200

200 Nguyen, T. X., S. V. Koppen, J. J. Ely, G. N. Szatkowski, J. J. Mielnik and M. T. P. Salud. 2007. Small aircraft RF interference path loss. In IEEE International Symposium on Electromagnetic Compatibility. EMC 2007 1 6. Nguyen, T. X., J. J. Ely, K. L. Dudley, S. A. Scearce, M. O. Hatfield and R. E. Ri. 2006. Passenger transmitters as a possible cause of aircraft fuel ignition in support of an aircraft accident investigation. In IEEE International Symposium on Electromagnetic Compatibility. EMC 2006 228 233. Nguyen, T. X., S. V. Koppen, J. J. Ely, R. A. Williams, L. J. Smith and M. T. P. Salud. 2004. Portable wireless device threat assessment for aircraft navigation radios. Nikitin, P. V. and K. V. S. Rao. 2009. LabVIEW based UHF RFID tag test and measurement system. IEEE Transactions on Industrial Electronics, 56(7): 2374 2381. Nikitin, P. V. and K. V. S. Rao. 2008. Antennas and propagation in UHF RFID systems. 2008 IEEE International Conference on RFID, 277 288. Nikitin, P. V. and K. V. S. Rao. 2006. Performance limitations of passive UHF RFID systems. In 2006 IEEE Antennas an d Propagation Socie ty International Symposium 1011 1014. Nikitin, P. V., K. V. S. Rao, R. Martinez and S. F. Lam. 2009. Sensitivity and impedance measurements of UHF RFID chips. IEEE Transactions on Microwave Theory and Techniques 57(5): 1297 1302. Niki tin, P. V., D. D. Arumugam, M. J. Chabalko, B. E. Henty and D. D. Stancil. 2010. Long range passive UHF RFID system using HVAC ducts. Proceedings of the IEEE PP(99): 1 7. Nordisk. 2010. Air cargo containers. Teleflex Incorporated. Available at: http://www .nordisk aviation.com/default.aspx?Cat=2&Id=11 Accessed 02/02 2010. NTIA. 2010. Federal spectrum use summary, 30 MHz 3000 GHz. NTIA. 2003. United States frequency allocations the radio spectrum. NTSB. 2010. NTSB Aviation Accident database & syn opses. National Transportation Safety Board. Available at: http://www.ntsb.gov/ntsb/query.asp Accessed 05/21 2010. Nunes, M. C. N., J. P. mond and J. K. Brecht. 2003. Quality of strawberries as affected by temperature abuse during ground, in flight and retail handling operations. In 239 246. Acta Hort. (ISHS).

PAGE 201

201 Nunes, M. C. N., J. P. mond K. V. Chau, M. Rauth, S. Dea and W. Pelletier. 2006. Effects of in store conditions on the quality of freshfruits and vegetables. 262. O'Connor, M. C. 2005. IATA app roves UHF for bag tags. Ott, R. L. and M. T. Longnecker. 2004. A first course in statistical methods. Belmont, CA, USA: Brooks/Cole Thomson Learning. Penttil, K., M. Keskilammi, L. Sydnheimo and M. Kivikoski. 2006. Radio frequency technology for automated manufacturing and logistics control. Part 2: RFID antenna utilisation in industrial applications. The International Journal of Advanced Manufacturin g Technology 31(1): 116 124. Poon, T. C., K. L. Choy, H. K. H. Chow, H. C. W. Lau, F. T. S. Chan and K. C. Ho. 2009. A RFID case based logistics resource management system for managing order picking operations in warehouses. Expert Systems with Applicatio ns 36(4): 8277 8301. Prothro, J. T., G. D. Durgin and J. D. Griffin. 2006. The effects of a metal ground plane on RFID tag antennas. In IEEE Antennas and Propagation Societ y International Symposium 2006 3241 3244. Rao, K. V. S. 1999. An overview of back scattered radio frequency identification system (RFID). In 1999 Asia Pacific Microwave Conference, 746 749 vol.3. Rappaport, T. S. 2002. Wireless communications : principles and practice. 2nd ed. Upper Saddle River, N.J.: Prentice Hall PTR. Rappaport, T. S and C. D. McGillem. 1989. UHF fading in factories. IEEE Journal on Selected Areas in Communications, 7(1): 40 48. Raumonen, P., L. Sydanheimo, L. Ukkonen, M. Keskilammi and M. Kivikoski. 2003. Folded dipole antenna near metal plate. IEEE Antennas and Pr opagation Socie ty International Symposium, 1848 851 vol.1. Reed, R. 2009. Refraction of light. Interactagram.com. Available at: http://interactagram.com/physics/optics/refraction/ Accessed 06/30 2010. Reitz, J. R., R. W. Christy and F. J. Milford. 1993. Monochromatic waves in bounded regions. In Foundations of electromagnetic theory, 441 469. ed. Anonymous Reading, Mass.: Addison Wesley. Roussos, G. and V. Kostakos. 2009. RFID in pervasive computing: State of the art and outlook. Pervasive and Mobile Computing 5(1): 110 131. Ruiz Garcia, L., P. Barreiro and J. I. Robla. 2008. Performance of ZigBee Based wireless sensor nodes for real time monitoring of fruit logistics. Journal of Food Engineering 87(3): 405 415.

PAGE 202

202 SAS Institute. SAS/STAT 9.1 user's gui de. Cary, N.C.: SAS Institute, Norwood, Mass. Saunders, B. 2003. Award winning blast resistant cargo container at Dubai air show spotlight. Schmoetzer, K. 2005. Enhanced cargo monitoring container communication interface. In International Aircraft Syst ems Fire Protection Working Group, Airbus, dept. BCECS4. Schmoetzer, K. 2001. Aircraft fire detection: Requirements, qualification and certification aspects. In 12th international conference on automatic fire detection, National Institute of Standards and Technology. Shahidi, R. 1995. Link budget analysis: student manual. Arlington, VA: LLC. Skorna, A. C. H. and A. Richter. 2007. RFID and air cargo. RFID Jounal. Available at: www.rfidjournal.com/article/articleview/3482/. 2008. Sydnheimo, L., L. Ukkonen and M. Kivikoski. 2006. Effects of size and shape of metallic objects on performance of passive radio frequency identification. The International Journal of Advanced Manufacturing Technology 30(9): 897 905. Syversen, E., F. J. Pineda and J. Watson. 2008. Temperature variations recorded during interinstitutional air shipments of laboratory mice. Journal of the American Association for Laboratory Animal Science 47(1): 31 36. Taylor, M. J. H. 1989. The aerospace chronology. London: Tri Service Press. The A luminum Association. 2008. Aluminum in aircraft. The Aluminum Association. Available at: http://www.aluminum.org/Content/NavigationMenu/TheIndustry/TransportationMar ket/Aircraft/default.htm Accessed 01/26 2010. Valenzuela, R. A., O. Landron and D. L. Jac obs. 1997. Estimating local mean signal strength of indoor multipath propagation. IEEE Transactions on Vehicular Technology, 46(1): 203 212. van de Voort, M. and A. Ligtvoet. 2006. Towards an RFID policy for Europe. DRR 4046 EC. Vega, H. 2008. Air cargo, trade and transportation costs of perishables and exotics from South America. Journal of Air Transport Management 14(6, Sp. Iss. SI): 324 328. Vronneau, S. and J. Roy. 2009. RFID benefits, costs, and possibilities: The economical analysis of RFID deploy ment in a cruise corporation global service supply chain. International Journal of Production Economics 122(2): 692 702.

PAGE 203

203 Villeneuve, S., M. O. Nagdi and J. P. mond 2001. Heat transfer in air cargo unit load devices. International Society for Horticultur al Science Acta Hortuculturae 566245 250. Villeneuve, S., F. Mercier, W. Pelletier, M. O. Nagdi and J. P. mond 2000. Effect of environmental conditions in air shipment of perishables during ground operations. In ASAE Annual International Meeting, ASAE. Villeneuve, S. 2006. tude des phnomnes d'change thermique et des conditions environmentales ayant un impact sur la conservation des denres prissables lors des oprations aroportuaires. PhD diss. Qubec: Universit Laval. Visich, J. K., S. Li, B. M. Khumawala and P. M. Reyes. 2009. Empirical evidence of RFID impacts on supply chain performance. International Journal of Operations & Production Management 29(11 12): 1290 1315. Wang, H., S. Chen and Y. Xie. 2010. An RFID based digital warehouse manag ement system in the tobacco industry: a case study. Internation Journal of Production Research 48(9): 2513 2548. Wegener, P. P. 1997. What makes airplanes fly? : history, science, and applications of aerodynamics. 2nd ed. New York: Springer. Wilmes, H., B Kolesnikov, A. Fink and C. Kindervater. 2002. New design concepts for a CFRP fuselage. Winder, S. and J. J. Carr. 2002. Radio and RF engineering pocket book. 3 rd ed. Oxford: Newnes. Wright, T. 2008. Temperature sensitive pharmaceutical transportation a changing picture. Wu, N. C., M. A. Nystrom, T. R. Lin and H. C. Yu. 2006. Challenges to global RFID adoption. Technovation 26(12): 1317 1323. Yonemoto, N., K. Yamamoto, K. Yamada and T. Hirata. 2007. RF emission measurement of 433 MHZ RFID tags for EM I evaluation to onboa rd instruments of aircrafts. In 7th International Symposium on Electromagnetic Compatibili ty and Electromagnetic Ecology, 232 235. Yuanxin, O., H. Yao, P. Ling, W. Dongwei and X. Zhang. 2008. An intelligent RFID reader and its applica tion in airport baggage handling system. In 4th International Conference on Wireless Communications, Networking and Mobile Computing, 2008. WiCOM '08. 1 4.

PAGE 204

204 Zhang Q. G. Cheng Z. Wang, J. Wei and D. Yan 2009. Development of RFID application system in car go inbound and outbound. In TENCON 2009 2009 IEEE Region 10 Conference, 1 6. Zweig, S. E. 2006. Advances in vaccine stability monitoring technology. Vaccine 24(33 34): 5977 5985.

PAGE 205

BIOGRAPHICAL SKETCH Magalie Laniel was born in Montreal, Quebec, Canada. She attended Laval University in Qubec city where she received, in 2004, a Bachelor of Engineering in food engineering. In t he following fall, she began a under the direction of Dr mond in the D epartment of Agricultural and Biological Engineeri ng at the University of Florida. work opened the path to continue her studies towards the PhD Along the years she has been working on several projects involving radio frequency identification (RFID) technology, packaging and transportation o f perishables.