<%BANNER%>

Development of Passenger Car Equivalency Values for Trucks at Signalized Intersections

Permanent Link: http://ufdc.ufl.edu/UFE0021137/00001

Material Information

Title: Development of Passenger Car Equivalency Values for Trucks at Signalized Intersections
Physical Description: 1 online resource (107 p.)
Language: english
Creator: Cruz-Casas, Carlos O
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: pce
Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Civil Engineering thesis, M.E.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Engineering DEVELOPMENT OF PASSENGER CAR EQUIVALENCY VALUES FOR TRUCKS AT SIGNALIZED INTERSECTIONS By Carlos O. Cruz-Casas August 2007 Chair: Scott S. Washburn Major: Civil Engineering Large trucks have considerably different size and performance characteristics than passenger cars. Consequently, these trucks can have a significant impact on traffic operations. It is therefore essential to properly account for this impact in the traffic operations analysis in order to reflect the operational quality of the roadway as accurately as possible. Signalized intersections are one roadway facility that can be particularly sensitive to the presence of commercial truck traffic. The most common method used for the analysis of signalized intersections is contained in the Highway Capacity Manual (HCM). In this method, the base saturation flow rate of the signalized intersection is defined in units of passenger cars per hour green per lane (pc/hg/ln). To account for the presence of large trucks in the traffic stream, the HCM includes a Passenger Car Equivalency (PCE) value. In the current edition of the HCM, a PCE value of 2.0 is applied for all large trucks, with no distinction between different sizes of trucks. Some transportation professionals have questioned the validity of this PCE value recommended by the HCM. They are concerned that the impact of trucks at signalized intersections is being under-estimated. If this is the case, then capacity is being over-estimated and intersections are not being adequately designed. The primary objective of this research was to determine appropriate truck PCE values to apply for signalized intersection analysis. These PCE values were classified by three different categories of truck sizes and performances. Additionally, a general PCE value with only one truck category was developed for planning purposes and/or a less detailed analysis. The development of the PCE values was based on the relative headway concept, as defined in the HCM. The results of this study are based primarily on data generated from a custom simulation program. However, a considerable amount of field data was collected for the purpose of simulation calibration. The PCE values determined from this study are 1.8, 2.2, and 2.8 for small trucks, medium trucks, and large trucks, respectively. Additionally, an equation was developed to calculate start-up lost time that accounts for the impact of trucks at the front of the queue, as opposed to the standard 2.0 seconds recommended by the HCM. Furthermore, based on the field data collected, it was found that the base saturation flow rate value of 1900 pc/hg/ln recommended by the HCM appears to be quite optimistic.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Carlos O Cruz-Casas.
Thesis: Thesis (M.E.)--University of Florida, 2007.
Local: Adviser: Washburn, Scott S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021137:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021137/00001

Material Information

Title: Development of Passenger Car Equivalency Values for Trucks at Signalized Intersections
Physical Description: 1 online resource (107 p.)
Language: english
Creator: Cruz-Casas, Carlos O
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: pce
Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Civil Engineering thesis, M.E.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Engineering DEVELOPMENT OF PASSENGER CAR EQUIVALENCY VALUES FOR TRUCKS AT SIGNALIZED INTERSECTIONS By Carlos O. Cruz-Casas August 2007 Chair: Scott S. Washburn Major: Civil Engineering Large trucks have considerably different size and performance characteristics than passenger cars. Consequently, these trucks can have a significant impact on traffic operations. It is therefore essential to properly account for this impact in the traffic operations analysis in order to reflect the operational quality of the roadway as accurately as possible. Signalized intersections are one roadway facility that can be particularly sensitive to the presence of commercial truck traffic. The most common method used for the analysis of signalized intersections is contained in the Highway Capacity Manual (HCM). In this method, the base saturation flow rate of the signalized intersection is defined in units of passenger cars per hour green per lane (pc/hg/ln). To account for the presence of large trucks in the traffic stream, the HCM includes a Passenger Car Equivalency (PCE) value. In the current edition of the HCM, a PCE value of 2.0 is applied for all large trucks, with no distinction between different sizes of trucks. Some transportation professionals have questioned the validity of this PCE value recommended by the HCM. They are concerned that the impact of trucks at signalized intersections is being under-estimated. If this is the case, then capacity is being over-estimated and intersections are not being adequately designed. The primary objective of this research was to determine appropriate truck PCE values to apply for signalized intersection analysis. These PCE values were classified by three different categories of truck sizes and performances. Additionally, a general PCE value with only one truck category was developed for planning purposes and/or a less detailed analysis. The development of the PCE values was based on the relative headway concept, as defined in the HCM. The results of this study are based primarily on data generated from a custom simulation program. However, a considerable amount of field data was collected for the purpose of simulation calibration. The PCE values determined from this study are 1.8, 2.2, and 2.8 for small trucks, medium trucks, and large trucks, respectively. Additionally, an equation was developed to calculate start-up lost time that accounts for the impact of trucks at the front of the queue, as opposed to the standard 2.0 seconds recommended by the HCM. Furthermore, based on the field data collected, it was found that the base saturation flow rate value of 1900 pc/hg/ln recommended by the HCM appears to be quite optimistic.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Carlos O Cruz-Casas.
Thesis: Thesis (M.E.)--University of Florida, 2007.
Local: Adviser: Washburn, Scott S.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021137:00001


This item has the following downloads:


Full Text





DEVELOPMENT OF PASSENGER CAR EQUIVALENCY VALUES
FOR TRUCKS AT SIGNALIZED INTERSECTIONS




















By

CARLOS O. CRUZ-CASAS


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ENGINEERING

UNIVERSITY OF FLORIDA

2007







































O 2007 Carlos O. Cruz-Casas



































To my parents, for their endless support









ACKNOWLEDGMENTS

I would like to acknowledge the excellent the guidance and support of the committee chair,

Dr. Scott S. Washburn, and committee members Dr. Lily Elefteriadou and Dr. Yafeng Yin. I

would also like to acknowledge Diego Arguea and Tom Hiles, for their significant collaboration

in this research. In addition, I would like to thank Matthew O'Brien, Darrell Schneider, and the

City of Gainesville for their support during the field data collection process.












TABLE OF CONTENTS


page

ACKNOWLEDGMENTS .............. ...............4.....


LIST OF TABLES ................. ...............7.__. .....


LIST OF FIGURES .............. ...............8.....


AB S TRAC T ............._. .......... ..............._ 10...


CHAPTER


1 INTRODUCTION ................. ...............12.......... ......


Background ................. ...............12.................
Problem Statement ................. ...............13.................
Research Obj ective and Tasks ................ ...............14........... ...
Document Organization............... ..............1

2 LITERATURE REVIEW ................. ...............16................


Overview of HCM Treatment of Heavy Vehicles ................. ...............16........... ..
Passenger Car Equivalency Factor at Signalized Intersections ................. ......................18
Car-Following M odels................ .... ..............2
Summary of HCM 2000 PCE Guidelines ................. ...............31...............
Summary of Previous PCE Studies ................. ...............31.......... ....
Summary of Car-Following Models ............... ...............32....

3 RESEARCH APPROACH ................. ...............34................


Method logical Approach .............. ...............3 4....
Field Data Collection ................. ...............37........... ....
Simulation M odeling .............. ...............49....

4 ANALYSIS OF FIELD DATA AND CALIBRATION OF SIMULATION MODEL.........56


Summary of Data Reduction ................. ...............56.......... ....
Time Head ways ...................... ...............57
Start-up Reaction Time (SRT) ................. ...............60................
Start-up Lost Time (SLT) .............. ...............62....
Calibration of Simulation Model .............. ...............63....
Execution of Experimental Design ................. ...............65................

5 ANALYSIS OF SIMULATION DATA .............. ...............67....


6 CONCLUSIONS AND RECOMMENDATIONS .............. ...............84....











APPENDIX


A DATA COLLECTION EQUIPMENT SETUP ................. ...............90........... ...

B PICTURES OF VEHICLE TYPES BY CATEGORY............... ...............92

C SIMULATION PROGRAM SCREEN SHOTS............... ..............9


D QUEUE COMPOSITION .............. ...............98....

E HEADWAY STATISTICS .............. ...............100....


F CALIBRATION RESULTS AND MSE CALCULATIONS .............. ....................10

LIST OF REFERENCE S ................. ...............105...__. ......


BIOGRAPHICAL SKETCH ....___ ................ ......._. ..........10











LIST OF TABLES


Table page


2-1 Car-following models comparison............... ...............3

3-1 Possible pairing combinations for four vehicle types* ................... ..............3

3-2 Data collection cites ................. ...............38................

3-3 Data collection periods for Williston Rd/34th Street site (method 1) .............. ................48

3-4 Data collection periods for other sites (method 2)............... ...............48...

4-1 Average headway and frequencies for each leader-follower combination ................... .....57

4-2 Time headways from field data............... ...............58..

4-3 Results from STATISTICA for model 1 .............. ...............58....

4-4 Results from STATISTICA for model 2 .............. ...............59....

4-5 Results from STATISTICA for model 3 .............. ...............59....

4-6 Results from STATISTICA for model 4 .............. ...............60....

4-7 Distribution of trucks by site and position in queue ................. ................ ......... .62

4-8 Final choice of calibration parameter values .............. ...............65....

5-1 Model estimation results for additional headways of 16 vehicle pair combinations .........73

5-2 Headways and PCE values for 16 vehicle pair combinations............... ..............7

5-3 Vehicle type's headway and their Ah .............. ...............76....

5-4 Time consumed and PCE values for each vehicle type ......... ................ ...............77

5-5 Results from STATISTICA for model with vehicle types .............. .....................7

5-6 PCE factors for three vehicle types............... ...............79.

5-7 Model estimation results for fHV equation with 16 vehicle pairs ................... ...............81

D-1 Vehicle type per position in queue............... ...............98.

E-1 Headway statistics for 16 lead-trail vehicle combinations .............. ....................10

F-1 Simulation model calibration results .............. ...............102....










LIST OF FIGURES


Figure page

2-1 Measured headways including start-up lost time (dark shade)............_ .. ........_.._. ....18

3-1 SW Williston Rd / SW 34th St aerial view (eastbound on Williston Rd) .........................38

3-2 SW Williston Rd / SW 34th St ground level view (eastbound on Williston Rd)..............39

3-3 University Ave. / Waldo Rd aerial view (northbound on Waldo Rd) .............. ................39

3-4 University Ave. / Waldo Rd ground level view (northbound on Waldo Rd) ....................40

3-5 US 41 / SR 50 aerial view (westbound on SR50) ................. ...............40...........

3-6 US 4 1 / SR 5 0 ground level view (westbound on SR5 0) ........._.._ ...... ._ ..............4 1

3-7 US 301 / SR 50 aerial view (eastbound on SR 50) ................. ...............41...........

3-8 US 301 / SR 50 ground level view (eastbound on SR 50)............... ..................4

3-9 CR 326 / SR 200A ground level view (westbound on CR 326) .............. ...................42

3-10 John Young Parkway / Colonial Dr aerial view (southbound on John Young Pkwy) ......43

3-11 John Young Pkwy / Colonial Dr ground level view (southbound on John Young
Pkw y) .............. ...............43....

3-12 Preferred video camera mounting location (plan view) for method 1 ............... ...............45

3-13 Screen capture (low resolution) of video image from Williston Rd/34th Street site in
Gainesville (method 1)............... ...............45...

3-14 Camera setup for data collection (method 2) ................. ...............46 ....._.__..

3-15 Screen capture of the composite video image (method 2) ........._.. ....... ._. ............47

3-16 Simulation program user interface ..........._ .....___ ...............52..

4-1 Start-up reaction time frequencies .............. ...............61....

4-2 Start-up reaction time frequencies for trimmed data set ....._____ ........___ ..............61

4-3 User-adj stable vehicle, driver, and model parameters ........................... ...............64

5-1 Queue sections for analysis............... ...............67

5-2 Impact of trucks in the first part of the queue to the saturation headway ................... .......68











5-3 SLT example ................. ...............72........... ....

5-4 Total time for vehicles 1-8 using vehicle pairs in positions 5-8 .............. ...................74

5-5 Time consumed by a large truck ................. ...............75........... ..

5-6 Total time for vehicles 1-8 using vehicle types in positions 5-8 .............. ....................78

A-1 Data collection equipment setup for method 1 .............. ...............90....

A-2 Signal controller cabinet with data collection equipment installed for method 1..............91

B-1 Small trcks. A) Panel truck. B) Garbage truck. C) Two-Axle Single-unit dump
truck. D) Small delivery truck. E) Passenger cars with trailers ................. ................. .92

B-2 Medium trucks. A) Three-Axle Single-unit dump truck. B) Concrete Mixer. C)
Passenger car with trailer using fifth wheel. D) Delivery truck. E) Single-unit cargo
truck. ............. ...............94.....

B-3 Large trucks. A) Tractor plus trailer. B) Tractor plus flatbed. C) Buses. .......................95

C-1 Simulation screenshot. A) Before signal turns green. B) Once the queue starts to
discharge .............. ...............97....









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Engineering

DEVELOPMENT OF PASSENGER CAR EQUIVALENCY VALUES
FOR TRUCKS AT SIGNALIZED INTERSECTIONS

By

Carlos O. Cruz-Casas

August 2007

Chair: Scott S. Washburn
Major: Civil Engineering

Large trucks have considerably different size and performance characteristics than

passenger cars. Consequently, these trucks can have a significant impact on traffic operations. It

is therefore essential to properly account for this impact in the traffic operations analysis in order

to reflect the operational quality of the roadway as accurately as possible. Signalized

intersections are one roadway facility that can be particularly sensitive to the presence of

commercial truck traffic.

The most common method used for the analysis of signalized intersections is contained in

the Highway Capacity Manual (HCM). In this method, the base saturation flow rate of the

signalized intersection is defined in units of passenger cars per hour green per lane (pc/hg/In).

To account for the presence of large trucks in the traffic stream, the HCM includes a Passenger

Car Equivalency (PCE) value. In the current edition of the HCM, a PCE value of 2.0 is applied

for all large trucks, with no distinction between different sizes of trucks.

Some transportation professionals have questioned the validity of this PCE value

recommended by the HCM. They are concerned that the impact of trucks at signalized

intersections is being under-estimated. If this is the case, then capacity is being over-estimated

and intersections are not being adequately designed.










The primary obj ective of this research was to determine appropriate truck PCE values to

apply for signalized intersection analysis. These PCE values were classified by three different

categories of truck sizes and performances. Additionally, a general PCE value with only one

truck category was developed for planning purposes and/or a less detailed analysis. The

development of the PCE values was based on the relative headway concept, as defined in the

HCM. The results of this study are based primarily on data generated from a custom simulation

program. However, a considerable amount of field data was collected for the purpose of

simulation calibration. The PCE values determined from this study are 1.8, 2.2, and 2.8 for small

trucks, medium trucks, and large trucks, respectively. Additionally, an equation was developed

to calculate start-up lost time that accounts for the impact of trucks at the front of the queue, as

opposed to the standard 2.0 seconds recommended by the HCM. Furthermore, based on the field

data collected, it was found that the base saturation flow rate value of 1900 pc/hg/In

recommended by the HCM appears to be quite optimistic.









CHAPTER 1
INTTRODUCTION

Background

Large trucks have considerably different size and performance characteristics than

passenger cars. Consequently, these trucks can have a significant impact on traffic operations. It

is therefore essential to properly account for this impact in the traffic operations analysis in order

to reflect the operational quality of the roadway as accurately as possible. This study focuses

only on those trucks that are considerably larger than pick-up trucks.

Signalized intersections are one roadway facility that can be particularly sensitive to the

presence of commercial truck traffic. Like other facilities, the length of trucks has a negative

impact on the capacity of the signalized intersection; however, the reduced performance

characteristics of these trucks has an even greater impact on signalized intersection than

uninterrupted flow facilities due to the need for many of the trucks traveling through an

intersection to decelerate to a stop and re-accelerate to cruise speed. Given the longer time it

takes trucks to re-accelerate to cruise speed, when compared to passenger cars, the presence of

trucks can have implications on signal coordination. This effect may be greater when trucks are

present at the front of a discharging queue.

When the traffic stream stops at the intersection, the inter-vehicle spacing decreases,

resulting in increased vehicle density. At this higher density, it is clear that trucks occupy more

space than passenger cars due to their physical characteristics. Once the signal turns green, and

after a short period of start-up lost time, vehicles start departing at the saturation flow rate and it

is here where the heavy vehicles have the greater impact due to their operational capabilities.

Large trucks have poorer acceleration than passenger cars; therefore it will take them more time

to reach their desired speed. With poorer acceleration characteristics, heavy vehicles slow down









the traffic stream and increase their time headway. Furthermore, with their longer length, heavy

vehicles increase the time headway of the vehicle following them. In arterial roads where the

Free Flow Speed is low, the cruising speed of all vehicles may be similar. Once the vehicles

reach their desired speed and have a constant cruising speed, the impact of trucks to the traffic

will be again mostly due to their length.

Problem Statement

The most common method used for the analysis of signalized intersections is contained in

the Highway Capacity Manual (HCM) [1]. In this method, the base saturation flow rate of the

signalized intersection is defined in units of passenger cars per hour green per lane (pc/hg/In).

To account for the presence of large trucks in the traffic stream, the HCM includes a Passenger

Car Equivalency (PCE) value. This factor is based on the relative headway of a truck to that of a

passenger car. In the current edition of the HCM, a single PCE value of 2.0 is used for all trucks

passing through a signalized intersection. Thus, this implies that a single large truck is

equivalent to two passenger cars for capacity analysis purposes; that is, an intersection can only

accommodate half as many trucks as cars.

Transportation professionals with the Florida Department of Transportation (FDOT) have

recently questioned the validity of this PCE value recommended by the HCM. This concern has

stemmed from observations that the failure rate of intersections in Florida seems particularly

high when any significant percentage of trucks is present in the traffic stream. Their

observations are especially disturbing in light of the fact that Florida' s population is growing

rapidly, and complimentary to this population growth is a corresponding growth in commercial

truck traffic. Thus, FDOT officials are concerned that the impacts of trucks at signalized

intersections in Florida are being under-estimated. If this is the case, then capacity is being over-

estimated and intersections are not being adequately designed.









Research Objective and Tasks

The primary obj ective of this research was to determine appropriate truck PCE values to

apply for signalized intersection analysis. These PCE values were classified by three different

categories of truck sizes and performances. Additionally, a general PCE value with only one

truck category was developed for planning purposes and/or a less detailed analysis. The results

of this study are based primarily on data generated from a custom simulation program.

However, a considerable amount of field data was collected for the purpose of simulation

calibration.

The following tasks were conducted to support the accomplishment of these obj ectives:

* Conduct a literature review
* Collect preliminary field data
* Develop data collection methods
* Determine appropriate criteria to select data collection sites
* Identify appropriate field data collection sites
* Collect video data from field
* Process the video using the RLRAP [2] to obtain signal status and time
* Reduce field data
* Develop a simulation program
* Calibrate the simulation program
* Develop an experimental design
* Generate simulation data set from experimental design
* Perform analysis and modeling of the generated simulation data set
* Develop new PCE factors

Document Organization

Chapter 2 presents an overview of the relevant studies found in the literature. This review

includes the state of the practice with regard to incorporating the effects of large vehicles into

traffic analyses, previous studies focused on the development of PCE values on arterial roads or

at intersections, and a review of car-following models that might be applicable to modeling

queue discharge at signalized intersections. Chapter 3 describes the research approach that was

used to accomplish the obj ectives of this study including the methodological approach, field data










collection, simulation model development, and the simulation experiments. Chapters 4 and 5

contain the analysis and results of the field and simulation data, respectively. Conclusions and

recommendations are contained in Chapter 6.









CHAPTER 2
LITERATURE REVIEW

An extensive literature review has been conducted in three areas. The first area is a review

of the state of the practice with regard to incorporating the effects of heavy vehicles into traffic

analyses, namely the Highway Capacity Manual [1, 3]. The second area targets previous studies

that focused on the development of PCE values at intersections. While other studies have been

done on the development of PCEs for other facilities, they were not covered in this proj ect due to

the unique influence of signals on traffic flow. A custom simulation program was developed for

use in this proj ect for several reasons, which are discussed in Chapter 3. Therefore the third area

of this chapter deals with a review of car-following models that might be applicable to modeling

queue discharge at signalized intersections.

Overview of HCM Treatment of Heavy Vehicles

The Highway Capacity Manual (HCM) first introduced the term "passenger car

equivalent" in the 1965 version of this publication [3] as "the number of passenger cars displaced

in the traffic flow by a truck or bus, under the prevailing roadway and traffic conditions." The

HCM 2000 [1] definition states that the passenger car equivalent is "the number of passenger

cars displaced by a single, heavy vehicle of a particular type under specified roadway, traffic,

and control conditions." Currently, a PCE value of 2.0 is specified for all heavy vehicles. The

refining of the original definition emphasizes the importance of traffic controls and their effect

on PCE values, and considers the subj activity of what can be considered as heavy vehicles by not

specifically citing those vehicles.

As defined by the HCM 2000, a heavy vehicle is any vehicle which has more than four

tires in contact with the driving surface. There is no distinction between trucks, recreational

vehicles, and buses in the calculation of the adjusted saturation flow rate at signalized









intersections. The HCM also recommends that if no data exist for a particular intersection, a

value of 2% heavy vehicles should be used for urban streets. The PCE is implemented through a

heavy-vehicle factor (fHV), which is used to adjust the base saturation flow rate. This heavy-

vehicle factor is one of several adjustment factors for the base saturation flow rate (So). In

Equation 2-1 is shown how the base saturation flow rate is adjusted by the as it is defined in

Equation 16-4 of the HCM 2000.

S = So x fHV [2-1]
The form of the equation for fHV (Equation 2-2) is:

f = [2-2]

Where :
PT = percentage of trucks in the traffic stream
ET = passenger car equivalency factor

The standard procedure for measuring saturation flow rate (HCM2000, 16-158 in appendix

H) prescribes that the headways of the first four to six vehicles in queue are not considered as

saturation headway because this time is usually considered to be part of the start-up lost time.

The HCM 2000 recommends a default start-up lost time (SLT) of 2.0 seconds if field

measurements are not available. It is not specified for what traffic stream composition this value

is based upon (e.g., passenger car only stream). This start up lost time can be determined by

adding the difference between saturation headway and the headway measured for those first few

vehicles that do not depart under the saturation headway.

Figure 2-1 illustrates this concept, where the saturation headway has a value of 2.0 seconds

per vehicle (light shade) and the lost time is shown in this example as applying to the first four

vehicles (dark shade).















~-H.-


35


Vehicle in Queue
Figure 2-1. Measured headways including start-up lost time (dark shade)

Passenger Car Equivalency Factor at Signalized Intersections

In 1987, Molina [4] derived a PCE model based on the assumption that passenger cars

depart at a constant saturation flow headway, and thus the basis for his model is the headway

method. Molina collected field data from one site in each of three cities in Texas, and obtained a

total of 13,000 observations. During the data collection, Molina considered vehicles to form a

part of the queue if they came to a complete, or near stop. He recorded the time that these

vehicles crossed the stop line. He classified the 13,000 vehicle observations into four vehicle

classes and within each vehicle class into ten queue positions. He used the regression analysis

method to develop a model of the collected data. An expression was derived for the additional

effect of a heavy vehicle in the first position of a queue. This is a modified expression of the

headway ratio method, and derived relationships consider only one heavy vehicle in the queue

with its position varying from one to ten.

In addition, the analysis was limited to through movements only, and other factors, such as

percentage of trucks, vehicular volumes, and headway increase of the eighth-positioned vehicle









behind the truck are not considered. Molina found that position in queue did not have a

pronounced effect when dealing with two- and three-axle and single-unit trucks, but had a very

pronounced effect with Hyve-axle combination trucks. His recommendations included using

different methods to distinguish between light and heavy vehicles when analyzing capacity at

signalized intersections, as these truck types can have significantly different effects.

Benekohal and Zhao [5] performed a study on the additional delay to the passenger cars

behind a heavy vehicle. This delay is produced from both longer headways as well as additional

headway increases of those vehicles behind the heavy vehicle that causes the delay. This study

introduces a new PCE value labeled the D-PCE, meaning a delay-based calculation of passenger

car equivalents. The delay-based passenger car equivalent is defined by Benekohal and Zhao as

the ratio of delay caused by a heavy vehicle to the delay of a car in an all-passenger car traffic

stream. The calculation of these D-PCE values considers the traffic volume as well as the

percentage of heavy vehicles in the traffic stream. Data were collected at ten approaches of

seven intersections in Central Illinois, where sites possessed as many ideal features as possible.

Vehicle headways, delays for all-passenger car streams, number of queued and non-queued

vehicles, the position of heavy vehicles in the queue, signal timing information, and geometric

data were all collected in the process. The headway time of the first vehicle was defined from

the moment the signal turned green to the point when the rear wheels of the vehicle crossed the

stop line. This implies that reaction time was included in the start-up lost time.

Because TRAF-NETSIM queue delay was used for comparison, queue delay was the

performance measure recorded in the field. Benekohal and Zhao concluded that the position of

the truck in queue is not as important as how many vehicles are behind the truck. Also,

comparison with the HCM values for single unit trucks indicates that the HCM overestimates the









effect by which capacity is reduced by these types of vehicles at signalized intersections. The D-

PCE increases with the number of vehicles behind a heavy vehicle, and PCE values for

signalized intersections should be determined based on additional delay caused by large trucks.

Kockelman and Shabih [6] performed a study of the impact of light-duty trucks (LDTs) on

the capacity of signalized intersections. Three factors were identified as influencing vehicle

headways: length, performance, and driver behavior. Kockelman and Shabih used the headway

method to arrive at PCE values for five different categories of light-duty trucks, where the

additional time it takes for a passenger car behind an LDT to enter the intersection relative to

being behind a passenger car is considered. The start-up lost time in their developed model did

not include the reaction time of the first driver, as measurements began only when the first

vehicle began to move (i.e., they did not record the start of green). In addition, only those

vehicles that came to a complete stop before the signal changed to green were considered to have

been part of the queue. Field data were collected from sites that met the following criteria:

* High traffic volumes and significant queuing
* Level terrain
* Exclusive left turn lane and protected signal phase for left turns
* Exclusive right turn lane
* Ease of data collection equipment setup
* Mix of vehicle types
* No parking zones along streets
* Insignificant disturbance from bus stops

The time elapsed from when the first vehicle in the queue began to move to when the rear

axle of the last vehicle in the queue crossed the stop line was measured. After a statistical

analysis, Kockelman and Shabih concluded that vehicle length is a significant factor on

following-vehicle headways. As a result, the impacts of LDTs should be given special

consideration, as sport-utility vehicles as well as vans have the ability to reduce the capacity at a

signalized intersection in a statistically significant manner. Kockelman and Shabih









recommended PCE values of 1.07, 1.41, 1.34, and 1.14 for small SUV, long SUV, vans and

pick-up trucks respectively.

Bonneson and Messer [7] performed a study in which several models were developed that

can be used to predict the saturation flow rate and start-up lost time of through movements at

signalized interchange ramp terminals and other closely spaced intersections. The minimum

discharge headway method is used, and it is typically reached by the vehicle in the sixth position

of the queue. In their models they include the term "traffic pressure", defined as the tendency for

vehicle headways to decrease as queue lengths increase and aggressive driving is present.

Bonneson and Messer indicate that other authors (Stokes et al.) have independently identified

this occurrence and call it "headway compression." Data for the study were collected at twelve

interchanges in five states. The sites selected for analysis contained the two basic forms of

interchanges, partial cloverleaf and diamond. Video cameras and computer-monitored tape

switch sensors, mounted at the upstream end of each of two street segments, were used for data

collection. The data were collected during weekdays, between 7:00 a.m. and 7:00 p.m.

The study concluded that there exists a strong correlation between start-up lost time and

saturation flow rate, and that the distance to a downstream queue as well as traffic pressure has a

significant effect on the saturation flow rate of a signalized traffic movement. Therefore, start-up

lost time is not a constant value as it is commonly used in practice, but rather dependent upon the

saturation flow rate. Bonneson and Messer recommended that an ideal saturation flow rate of

2000 pc/h/In should be used for high volume intersections in urban areas.

Bonneson et al. [8] studied the formulation of the equation 16-4 of the HCM 2000 for

saturation flow rate at intersections. The research resulted in the development of some new

adjustment factors such as area population, number of lanes, and right-turn radius and the









revision of some existing factors such as right turn, traffic pressure, and heavy vehicle. These

adjustment factors were developed from data collected five different counties in Florida. Data

were collected at 12 intersections, which included a total of 38 approaches and 2901 cycles in

overall.

While not a focus of this study, a truck PCE factor was included in the model, as some

trucks were present in the data set. Their estimated PCE value was 1.74, which interestingly is

smaller than the HCM recommended value of 2.0. However, the authors indicate that truck

percentages were not significant in their data set, and they specifically recommend a more

thorough investigation of this specific factor.

Perez-Cartagena and Tarko [9] in their study focused in the development of local values of

the base saturation flow rate and lost times used in capacity analysis of signalized intersections in

Indiana. In their study, it was considered that the first four vehicles in queue were carrying the

SLT. They found that some of the default values recommended by the HCM were adequate for

their location. These factors included the heavy vehicle factor @~v). They also found that the

saturation flow rate was not the same for all their sites even though they were almost identical in

terms of geometry and traffic conditions. This indicated that there are some other factors

affecting the saturation flow rate that are not considered in the HCM. Perez-Cartagena and

Tarko proposed population adjustment factors of 0.92 for medium towns and 0.79 for small

towns to be added to the adjusted saturation flow equation 16-4 in HCM 2000.

Li and Prevedouros [10] examined saturation headway and start-up lost times of traffic

discharging from a signalized intersection. Their study was done using data collected from one

through movement and one protected left turn at a single intersection. They proved that the

assumption of that the saturation headway of short and long queues is the same was overlooking









other factors that might be present. It was observed how the last few vehicles in a longer queue

can produce either compressed or elongated headways.

Compressed headways were observed when vehicles bunch together to be able to cross the

stop bar before the clearance interval is over. Elongated headways were observed when the

queues were long enough to allow the vehicles to exceed speeds of 40 mph. Additionally, they

found that the minimum headway was not reached until the 9th to 12th vehicle in queue. In

addition Li and Prevedouros recommended a mean start-up reaction time of 1.76 seconds with a

standard deviation of 0.61.

Car-Following Models

Cohen investigated the issue of simulating queue discharge at a signalized intersection

through the application of the modified Pitt car-following model [l l]. This study was based on a

single intersection with no restrictions on the departing flow. Some queue-discharge

mechanisms are based on the assumption of every vehicle in queue departing from the

intersection at equal time headways. These assumptions are neglecting the start-up delay

brought from the first few cars in queue and other issues such as varying vehicle and driver

characteristics, among others.

The basic form of the modified Pitt car-following model is shown in Equation 2-3.



+x v, (t + R) vt (t + R)- xT hxv
+[vfl~l)v,(+Xlxl. 2 xaz(t +R)x T"
af (t + T) = [2-3]

I~~t2xT1

This model estimates an acceleration for a following vehicle, subj ect to three constraints: af

has to be between as,,, and a;;;; the speed at any time t has to be less than the free-flow speed;

and af has to be less than the acceleration computed for safe following (Equation 2-4):










-0.5xy2(t+R)2
a ye (t + T) = f T\2 [2-4]
s, (t + T) sf (t + T) L, i31ll~ im
2 xatmmn

In these two equations the variables are defined as:
a,(x) = acceleration trailing vehicle at time x, computed from car-following
(ft/s2)
t = current simulation time (sec)
T = simulation time-scan interval (sec)
K = sensitivity parameter used in modified Pitt car-following model
sl(x) = position of lead vehicle at time x as measured from upstream (ft)
R = perception-reaction time (assumed to be equal for all vehicles) (sec)
sf(x) = position of follower vehicle at time x as measured from upstream (ft)
L1 = length of lead vehicle plus a buffer based on jam density (ft)
h = time headway parameter in Pitt car-following model (buffer headway)
(sec)
vf(x) = speed of follower vehicle at time x (ft/s)
vl(x) = speed of lead vehicle at time x (ft/s)
al(x) = acceleration of lead vehicle at time x (ft/s2)
ap, = acceleration of follower vehicle as computed from safe following
(ft/s2)
vlmzn = minimum speed of lead vehicle (ft/s)
almzn(vt) = minimum acceleration of lead vehicle (maximum deceleration) (ft/s2)

For purposes of the research, the initial movement of the vehicles was defined once they

reached the speed of one foot per second. Furthermore, it was found that an appropriate value

for the parameter K should be greater than 1. This is taking into consideration the assumption

that a driver in a queue behaves significantly different than a driver in free-flowing traffic. The

best fit for this value in the study was 1.25.

Greenshields [12] indicates that the vehicle headways become fairly constant after the fifth

vehicle. This indicates that the first four cars should not be used for the discharge headway

calculations. These cars are the major carriers of the start-up lost time. The results of this

research include only passenger cars.

In the study it was found, as expected, that longer vehicle length results in longer discharge

headways. Furthermore, it was found that the expansion wave speed is independent of the Free









Flow Speed (FFS) but it increases when the following distances are closer. The expansion wave

may slow down when the vehicles' average speed reaches 30 mph. This is reasonable since

acceleration decreases at higher speeds.

This research indicates that the impact of a truck is greater if the truck is positioned in the

first few positions. Additionally, it is recommended that the K parameter should be calibrated

with vehicles in the second and third position since they are the vehicles most affected.

Bonneson [13] studied and summarized previous researches of modeling discharge

headways at signalized intersections. Bonneson selected a model based on vehicle and driver

capabilities, including driver reaction time, driver acceleration, and vehicle speed. Then, he used

field data from five different sites to calibrate the model.

The Briggs headway model [14] assumes that the acceleration for each vehicle in queue

remains constant. This model separates the vehicles not by position but by distance to the stop

bar in comparison with the distance they need to reach their desired speed. This distance is

denoted as d in Equation 2-5.



nxd max [2-5]
Then

2ixdxn 2~xdx(n-1)
h = T+-
"A A [2-6]
Otherwise


h,= T +-
[2-7]
With



m 2 xa [2-8]









Where :
n = queue position (n = 1, 2, 3...)
d' = distance between vehicles in a stopped queue (ft)
danax = distance traveled to reach speed V, (ft)
h,, = headway of the nth queued vehicle (sec)
T = driver starting response time (sec)
a = constant acceleration of queued vehicles
V, = desired speed of queued traffic (ft/s)

Messer and Fambro [15] found that regardless of the position of the driver, the response

time was 1.0 sec. Although it was found that an additional delay of 2.0 sec should be allocated

to the first driver. They also found that an average length for each queue position is 25 ft.

To these values it is important to add the additional length and delay due the vehicle type.

In this report the vehicle composition of the queue is not specified.

Buhr et al. [16] conducted a study of driver acceleration characteristics on freeway ramps.

This study considers only passenger cars from a stopped condition. The authors determined that

acceleration decreased linearly with increasing speed according to Equation 2-9.



;Vmax y [2-9]
Where :
a = instantaneous acceleration (ft/s2)
Ana = maximum acceleration (ft/s2)
V = velocity of vehicle (ft/s)
Y;,a = maximum speed corresponding to zero acceleration (ft/s)

For on-ramps at level terrain, the authors found an Ana = 15 ft/s2and y;2a = 60 ft/s. A later

study conducted by Evans and Rothery [17], studied the acceleration and speed characteristics of

queued drivers. The results show an exponentially increasing speed to a desired speed of

approximately 50 ft/s. This speed is reached by the vehicle in 25th pOSition, which can be

considered as normal cruising speed when crossing the stop bar.









Previous studies indicate that the discharge headway (h) between the nth vehicle and the

(n-1) th vehicle has two components. First, the time that it takes from the beginning of the green

time to the driver to start moving. This can be estimated as r + n x T, where r is the additional

time for the first driver and T is the individual response time. The second part is the time that it

takes the vehicle to reach the stop bar. Briggs assumed that each vehicle in queue will occupy

the same amount of space. The final form of the proposed headway model is shown in Equation

2-10.

s y t(n) s t(n-- 1)
hn = rx N, + T + +
V, a,
maxmax [2-10]
Where :
h,, = headway of the nth queued vehicles (sec)
r = additional response time of the first queued driver (sec)
1 if it is the first vehicle
0,= otherwise
T = driver starting response time (sec)
d = distance between vehicles in a stopped queue (ft)
Vm,; = maximum speed (ft/s)
Vsr(n)= stop line speed of the nth queued vehicle (ft/s)
A,,; = maximum acceleration (ft/s2)

This approach does not consider any kind of mixed traffic. Another limitation in this

model is that it does not consider any variation in the inter-vehicle length or any other buffer

length between the two vehicles.

In conclusion, this study says that the minimum discharge headway is dependent on driver

response time, desired speed and traffic pressure for each movement. It states that the minimum

discharge headway is not reached until the eighth vehicle. This model also suggests that under

ideal conditions the discharge headway should be shorter than 2.0 sec/veh.

Akgelik et al. [19] describe an exponential queue discharge flow and speed model. In this

study they model queue discharge speed in addition to the headway. By including speed it is









easier to develop relationships for traffic parameters like vehicle spacing, density, time and space

occupancy ratios, gap time, occupancy time, space time and acceleration characteristics.

Equations 2-11, 2-12, and 2-13 show the models for speed, flow and headway developed

by Akgelik, et al [19]:

v, v:,[1- e "'""]
s [2-11]
qs = q,[1 e-;nr,(-ty[212



[2-13]
Where :
v, = queue discharge speed at time t (kin, b)
v,, = maximum queue discharge speed (km/h)
m, = parameter
t = time since the start of green (seconds)
tr = start response time (constant) average from all drivers in Is' position
qs = queue discharge flow rate at time t (veh/h)
q,, = maximum queue discharge flow rate (veh/h)
m, = parameter
h, = queue discharge headway at time t (seconds)
h, = 3600/qs
h,, = minimum queue discharge headway (seconds)
h, =3600/q,,

Pipes [20] describes the ideal space or distance headway as one car length for every ten

miles per hour of speed at which the follower vehicle is traveling. The resulting equation is

shown in Equation 2-14.


d mm, = [x,(t) t = L" (1 47)( tl)0) + L) [2-14]


Where :
d,,,,, = distance between vehicles at time t (ft)
x,, = location of lead vehicle at time t (ft)
x,, 3 = location of follower vehicle at time t (ft)
L,, = length of lead vehicle (ft)
v,, 3 = speed of the following vehicle at time t (mph)









In this case the headway depends more on the length of the lead vehicle. For example, say

a passenger car is following a truck, the ideal distance will be using the lead vehicle's length. In

this case it will result in an extremely large headway. On the other hand, if a truck is following a

passenger car, the model gives a smaller headway. It is obvious this model is not taking in

consideration the driver behavior and the braking capabilities of the vehicles. Previous studies

have found that it takes longer to a truck to stop. Therefore, the model is not compatible with

real life scenarios.

This model is relatively easy to use, the mathematical operations are simple. The big

limitation of this model is that it does not take into consideration any kind of interaction between

vehicles. The acceleration is considered to be constant and there is no way to allocate heavy

vehicles.

As appear in May [21], Forbes' theory approaches the car-following by considering the

reaction time of the follower. The minimum gap between the vehicles should be greater than the

reaction time of the drivers. In addition to the minimum gap, this model considered the length of

the vehicles. In Equation 2-15 is expressed this mathematical relationship.


h y = At +
v (t) [2-15]

Where :
h = time headway (seconds)
At = reaction time, assumed to be 1.5 sec (seconds)
Ln = length of lead vehicle (ft)
vn = speed of the lead vehicle at time t (mph)

May [21] also presents the studies developed by a group of researchers from General

Motors about car-following theories. These studies were more extensive than the studies made

for Pipes' model and Forbes' model. It also has a particular importance because it is based on an

empirical model. After four previous intents or versions, GM came up with the 5th and final









model. This eliminates the discontinuities in the previous versions. This final model is shown in

Equation 2-16.

a[v, (t + At)l"
an (t + At) = 2, n()-v(t)l
[x, (t) x1 (t)' [2-16

Where :
an+1(x) = acceleration of follower vehicle at time x (ft/s2)
At = reaction time (seconds)
a = sensitivity parameter with speed and distance exponents m and I
respectively
m = speed exponent
1 = distance exponent
vn+l(x) = speed of follower vehicle at time x (ft/s)
vn(x) = speed of lead vehicle at time x (ft/s)
xn(x) = position of lead vehicle at time x as measured from upstream (ft)
xn+l(x) = position of follower vehicle at time x as measured from upstream (ft)

This model includes the acceleration characteristics of the following vehicle and yet it does

not include a variation for vehicle type or size. However, the parameters m and I can be

calibrated by individual vehicle types to adjust the acceleration of the following vehicle.

Another limitation of this model is the lack of consideration for the vehicle' s length.

Long [22] found that passenger cars, SUVs, and vans length averaged around 15 ft. These

vehicles length vary from 10.9 to 19.7 ft, with approximately two-thirds of the sample range

between 13 and 16 ft. It was also found that when any of these cars were pulling a trailer, their

total length increase in average to 35 ft.

The truck distribution was not as close to a normal distribution as the PC's were.

Additionally, only an approximated 12 % of the trucks were as long as the WB-50 design length

(55 ft) or shorter. Long [22] found that combination trucks typically have a length of 65 ft.

Using the grouping that AASHTO has follow for years regarding acceleration capabilities,

typical lengths of 15, 65 and 30 ft can be used for PCs, combinations trucks and all others









vehicle respectively. This research also specifies some other types of trucks that can potentially

be used as guidance.

Then an expected average vehicle length (EVL) could be estimated with weighted average

using the expected proportion for each one of the groups. Long also recommend a reasonable

inter vehicle spacing of 12 ft in contrast to the 3 ft default value in CORSIM.

Summary of HCM 2000 PCE Guidelines

* Currently, a PCE value of 2.0 is specified for all large vehicles.

* No distinction is made between trucks, recreational vehicles, and buses in the calculation
of the adjusted saturation flow rate at signalized intersections. It applies to any vehicle
with more four tires in contact with the driving surface.

* The headways of the first four to six vehicles in queue are not considered as part of the
saturation headway.

* A default start-up lost time (SLT) of 2.0 seconds, if field measurements are not available,
is recommended.

Summary of Previous PCE Studies

There are no recent studies that have examined or determined heavy vehicle PCE values

for signalized intersections. There are previous studies that demonstrate how heavy vehicles

have an impact on traffic streams at an intersection. Of all the research reviewed, headway was

by far the most common performance measures used to base PCE values on. Significant Eindings

from these studies are as follows:

* Molina derived a different expression for the additional effect of a heavy vehicle in the
first position of a queue. Molina also found that position in queue did not have a
pronounced effect on two- and three- axle and single-unit trucks.

* Benekohal and Zhao concluded that other than being in the first position, the position of
the heavy vehicle in the rest of the queue does not matter, but yet the number of cars
behind the heavy vehicle has a significant impact.

* Benekohal and Zhao concluded that at signalized intersections, the additional delay caused
by larger trucks should be used for calculating PCEs. They introduced a new term, D-
PCE, meaning a delay-based calculation for PCE.










* Kockleman and Shabih found that three factors influence the vehicle headways: length,
performance and driver behavior. In addition they recommended PCE values of 1.07,
1.41, 1.34, and 1.14 for small SUV, long SUV, vans and pick-up trucks respectively.

Bonneson et al. developed new and revised adjustments to the saturation flow equation.
They estimated a PCE value 1.74 for heavy vehicles, but that was based on a data set with
minimal truck observations.

Perez-Cartagena and Tarko stated that the fHV is adequate for use in Indiana, although they
recommend an additional factor for population.

Li and Prevedouros found that the minimum headway is not reached until the 9th vehicle
crosses the stop bar. They also recommended a mean start-up reaction time of 1.76
seconds with a standard deviation of 0.61.

Summary of Car-Following Models

There were several car-following models reviewed in this chapter including from the first

few models developed to the most recent ones. Table 2-1 summarizes the details that were taken

into consideration at the moment of selecting which car-following model was going to be used.

Table 2-1. Car-following models comparison
Pipes Forbes GM Modified Pitt
Constant acceleration Yes Yes No No
Acceleration of the leading vehicle No No No Yes
Speed at stop bar No No Yes Yes
Following distance Fixed through vehicle type Fixed Fixed Fiecutanbhdsedwh
Vehicle length Yes Yes No Yes
Length variation for vehicle type No but can be adjusted No but can be adjusted No Yes but have to be adjusted
Calibration difficulty High High High Medium
Implementation difficulty Low Low Low Low
Computational efficiency High High Medium Medium

Of these five car-following models studied, the modified Pitt model fits better to apply the

model to a signalized intersection with heavy vehicles in the stream. In the model, the

summation of the length of each vehicle plus a buffer length determined at j am density is known

as L. For passenger cars, a value of 20 ft is used for this length L. This value should be

reconsidered due the findings in Long's study [22]. This value should be the sum of two parts.










The first part should be the length of the leading vehicle, and second the minimum or allowable

inter-vehicle spacing. By doing this, a more detailed composition of the traffic stream can be

reached.









CHAPTER 3
RESEARCH APPROACH

This chapter describes the research approach that was used to accomplish the obj ectives of

this study. More specifically, it will discuss the methodological approach, field data collection,

simulation model development, and the simulation experiments.

Methodological Approach

As discussed in the literature review chapter, two different methodological approaches

have been used in previous PCE related studies. One was based on the concept of a delay-based

passenger car equivalent for trucks. The other was based on the concept of a time headway

passenger car equivalent. The latter is the approach that is currently used in the Highway

Capacity Manual [1]. One of the main objectives of this project was to update the PCE values

used for large trucks in FDOT's ARTPLAN software [23]. Since the ARTPLAN calculation

methodologies are largely based on the HCM analysis framework, it was decided to use the time

headway approach for purposes of consistency.

In the headway based approach to PCE calculation, it is important to recognize that the

time headway value between successive vehicles is a function of both the leading vehicle and the

trailing vehicle. Thus, a PCE value that is determined for a particular type of vehicle must

account for a variety of different vehicles that may precede it in a queue. For example, the

headway between a passenger car (leader) and large truck (follower) will be different than that

between a large truck and a large truck, all else being equal. Of course, it is not reasonable to

expect practitioners to collect the exact sequence of vehicles in queue during data collection

activities. Generally, the data collection is limited to just a percentage of trucks in the traffic

stream, not their actual positions in the traffic stream as well. Furthermore, this sequence will

usually be different from cycle-to-cycle. Thus, the PCE values that are derived must be









somewhat generalized so that the sequence of vehicles is not an input to the selection of a PCE

value.

While it might seem desirable to develop the PCE values based on a large number of the

different leader-follower vehicle combinations, there are some practical concerns with this

approach. First, certain combinations of vehicles will occur very rarely in the traffic stream; for

example, a recreational vehicle pulling a trailer that is following a motorcycle, or vice versa, as

the individual frequencies of occurrence for these vehicle types is quite small. Second, from a

data collection perspective again, it is only reasonable to expect practitioners to classify vehicles

into just a few different categories. Therefore, for this proj ect, it was decided to consider just

three different truck types, in addition to the passenger vehicle.

Different types of trucks have different impacts on the traffic stream; for example, one

single-unit truck does not take up the same amount of space as one tractor+semi-trailer, and their

acceleration capabilities are most likely different as well. Therefore, heavy vehicles were

categorized as either a small, medium or large truck depending on their size and operational

characteristics. The details of each category will be described in a later section in this chapter.

The end result of this proj ect is the development of three different PCE values applicable

to three categories of heavy vehicles which can be used to calculate a heavy vehicle factor (fHV)

analogous to the current HCM method, as shown in Equation 3-1.


f,
HV Ps, x (E, -1)+P x (E, -1)+P,, x (E,, -1)) [3-1]

Where :
fHV = adjustment factor for heavy vehicles in traffic stream
P, = proportion of truck type i in traffic stream
E, = PCE factor for truck type i
i = LT for large truck, M~T for medium truck, and ST for small truck










According to the procedure in the HCM to obtain the saturation flow rate, it was necessary

to measure the saturation headway at each intersection with at least eight vehicles in queue.

Since it was established that the headway of each vehicle depends on the leading and trailing

vehicle and that there were three different truck types under study, the number of possible leader-

follower combinations is 16 (42). Table 3-1 enumerates these 16 combinations.

Table 3-1. Possible pairing combinations for four vehicle types*
PC-PC ST-PC MCT-PC LT-PC
PC -ST ST -ST MCT -ST LT -ST
PC -MT ST-M(T M~T-MT LT-M(T
PC-LT ST-LT MCT-LT LT-LT
* PC = passenger car, ST= small truck, M~T= medium truck, LT= large truck

Since the focus of this study was on truck PCE values, it was obviously necessary to find

sites with a relatively high percentage of large trucks. For sites with less than 10% trucks in the

traffic stream, a very large percentage of the queues would have only zero or one truck in it.

Thus, the number of cycles that would need to be collected to obtain a significant number of

queues with multiple trucks present would be extreme, thus leading to a very lengthy and

inefficient data collection process. Thus, only sites with a truck percentage of at least 10% were

considered for data collection. However, sites with persistent queues of at least 8 vehicles and

10% or more large trucks are not very common. Sites with a lot of traffic (thus generating the

necessary queue lengths) typically have small truck percentages, and sites with high truck

percentages (such as along truck routes) do not typically have high overall volumes. With the

data collection resources available for this proj ect (i.e., time, money, labor), it was not feasible to

obtain enough data from enough sites to facilitate the obj ectives of this proj ect from field data

alone. Therefore, the decision was made to collect as much field data as project resources would

allow, and then use these field data to calibrate a simulation model that would be used to provide

the full data set upon which to base the development of the truck PCEs. The collection of the










field data is described in the next section, and the development and application of the simulation

model in the section after that.

Field Data Collection

In order to reduce the impact of additional factors that could affect the effectiveness of the

intersection, and to obtain the necessary amount of data, it was essential for the sites selected to

contain certain characteristics. The criteria to select these sites for field data collection are

outlined below.

Site Selection Criteria

Geometry

* Typical four leg intersection (turning radii at or close to 900)

* At least one site with only one through lane. The other sites should have two or three
travel lanes in the through direction

* The approach must have an exclusive left tumn lane(s), but the queue cannot spill back onto
through lanes

* Sites with an exclusive right turn lane are preferred, but not absolutely necessary if the site
has a small percentage of right turning vehicles

* Level terrain is preferred, but sites with small grades are acceptable

* No curbside parking or bus stops near the intersection, or other external factors that will
significantly influence the saturation flow rate


Traffic

* Queue lengths of at least 10 veh/lane at the beginning of green should be regularly present
during the data collection period. This is a function of the signal g/C ratio and cycle
length, in addition to the overall traffic demand. For example, for a g/C ratio of 0.4 and a
cycle length of 60 seconds, an average hourly flow rate of at least 1000 veh/hr/lane is
needed. Additionally, if progression is favorable, then the volumes would need to be
adjusted upward slightly, as a higher percentage of vehicles would be arriving on green.

* A heavy vehicle percentage of 10% or higher in the traffic stream.












*The operations of the observed approach must not be impacted by a downstream queue.
That is, vehicles must be able to depart freely from the subj ect approach.


From these criteria, and assistance from FDOT personnel, six sites were ultimately chosen


for field data collection. These sites are listed in Table 3-2.


Table 3-2. Data collection cites
Site # Intersection street names
1 SW Williston Rd / SW 34" St
2 Waldo Rd / University Ave
3 US 41 / SR 50
4 US 301 /SR 50
5 SR 326 /CR 200A
6 John Young Parkway / Colonial Drive


City
Gainesville
Gainesville
Brooksville
Brooksville
Ocala
Orlando


Sites Selected


Herein are shown aerials and ground level views from each site selected.


is







s ~e "r
4.
k~si
n


ii
re

c


I


/I


Figure 3-1. SW Williston Rd / SW 34th St aerial view (eastbound on Williston Rd)









































Vr 34th St ground level view (eastbound on Williston Rd)


lrr~7
?tl


~ rl. -r.--
ul


Figure 3-3. University Ave. / Waldo Rd aerial view (northbound on Waldo Rd)


~;FL n
Noiil1
jl r,
I~Lr ~
1 AR



































--rrr
~~-----~'


iv
~r~7c ;~"~ '

;ur~
-":~JF"
ii" l~~m~
.;1
I~





Figure 3-4. University Ave. / Waldo Rd ground level view (northbound on


winkrdid5 Camrhi ii) 1!992-2100 ES~RI Irs:

Figure 3-5. US 41 / SR 50 aerial view (westbound on SR50)


































ground level view (westbound on


Figure 3-7. US 301 / SR 50 aerial view (eastbound on SR 50)





































Figure 3-rr. UY JUl / YK SU ground level view (eastbound on YK SU)


Figure 3-9. UK 320 / YK ZUUA ground level view (westbound on UK 320)





,~,.-cl*~I- .rL~~LL*I~-' ~--------~
~I ~aar r-
...,.... -- -1-
....~ eL 1I~L I
i


~sr
I~I~---- .
'4EL~"

~ ur~c
r~mrl- ------; -,;
I I -
~~L~ylruurl~~ *~nr~;.
t ;~r/w~FYrCil*Wlfl;lr;** k-1~
r'


John Young Parkway / Colonial Dr aerial view (southbound on John


Figure 3-10.


Young Pkwy)


Figure 3-11. John Young Pkwy / Colonial Dr ground level view (southbound on John Young
Pkwy)


Noiil1
..: .e .,~n~~ ~ikiil


r i









Data Collection Methods

Two different methods for data collection were used. One method used a single camera

along with equipment installed in the signal controller cabinet to obtain signal status information

concurrent with the video signal. This method is referred to as "method 1," and was used for

only the Williston Rd/34th St site. The other method used two cameras to obtain both traffic and

signal status information, and is referred to as "method 2." Each of these methods is described in

the following sections.

Data collection equipment for method 1

For this method, it was necessary to be able to set up the following devices inside of the

signal controller cabinet: a VCR, a signal encoding device, and current sensors that attach

(passively) to each of the green bulb/LED power wires. Figure A-1 illustrates the necessary

connections. A picture of the installation of the equipment in a signal controller cabinet is shown

in Figure A-2.

A video camera was mounted on the mast arm facing the approach of interest. An

additional constraint for this setup was that the camera be mounted on a mast arm that is in the

same quadrant as the signal controller cabinet to easily facilitate running the video/power cable

from the video camera to the controller cabinet. Figure 3-12 illustrates the camera setup at the

Williston Rd/34th St intersection.

The stop bar and back of queue must be visible within the camera field-of-view (FOV), as

demonstrated in Figure 3-13. Figure 3-13 also shows how the video image and the signal status

information are combined into a composite view to facilitate data reduction. This is

accomplished through special processing hardware and software that is too detailed to explain

here. For more information on this system, see Washburn et al. [2].





1
I` 1







I I
I L


I~


Figure 3-12. Preferred video camera mounting location (plan view) for method 1
Ash a~SgtII'B
i" f tF ~ i* ir ilrr~~ rrum


Figure 3-13. Screen capture (low resolution) of video image from Williston Rd/34th Street site
in Gainesville (method 1)


.1 I~ L









Data collection equipment for method 2

To be able to implement the first method for data collection requires the cooperation of the

local agency that maintains the signals. Due to an existing relationship with the City of

Gainesville, this cooperation was easily facilitated. However, it was not so easily facilitated in

the other jurisdictions. Thus, it was decided to use a different data collection approach for sites

outside of Gainesville to avoid this complication. This second method requires no access to the

control cabinet or the need of the presence of local agency staff. This method consists of having

two cameras at the site at ground level. One camera is placed in a position where it can focus on

the traffic signal head and at the same time see a fair amount of the queue. The second camera is

placed where it has a FOV that includes the stop bar and the first couple of cars in the queue.

This setup is illustrated in Figure 3-14.


Figure 3-14. Camera setup for data collection (method 2)










For this method, if the back of queue was not visible within the camera FOV for some

cycles, it was recorded manually. Method 2 was used for data collection at the other five sites.

While one of the other sites was in Gainesville, method 1 could not be used at this site due to

complications of running the necessary cables from the video camera through the in-ground

conduit to the signal controller cabinet.

To facilitate data reduction with this method, the two camera views had to be combined

into a single composite image, along with a timer, as shown in Figure 3-15. This was

accomplished with a video mixer, a software program that generated the timer, and another

software program that merged the timer with the output of the two camera images from the

mixer.


Figure 3-15. Screen capture of the composite video image (method 2)











Data Collection Periods


For the data collection performed at Williston Rd/34th St site, there was more flexibility in


how much data could be collected, and at what times data could be collected. This was because


with method 1, the VCR in the controller cabinet could be programmed to record any time


period, and the City of Gainesville allowed the research team to access the cabinet to change


VCR tapes at our discretion. For data collection with method 2, it was necessary to be present on


site during the entire duration of the data collection. Thus, due to this issue and travel costs and


constraints, data collection was limited to a 4-hr period at the sites outside of Gainesville.


The data collection periods for the Williston Rd/34th St site (using method 1) are shown in


Table 3-3.


Table 3-3. Data collection periods for Williston Rd/34th Street site (method 1)

Tape # Date Time(s)

2 3 2 2006 6:30 9:30 am; 11:30 am 1:30 pm; 3:30 6:30 pm

5 3 31 2006 8:30 am 4:30 pm

6 4 12 2006 6:30 am 2:30 pm


The data collection periods for the sites using method 2


Table 3-4. Data collection periods for other sites (method 2)
Site City View' Lanes Tape num Time interval Toatie
hr:min
1 Gainesville Stop bar 2 1 3:30 -5:30 pm 2
2 Ocala Stop bar 1 1 8:00 -12:00 pm 4
Green 2 8:00 -12:00 pm 4
3 Orlando Stop bar 3 1 8:00 -12:00 pm 4
Green 2 8:00 -12:00 pm 4
1 Gainesville Stop bar 2 1 2:00 -6:00 pm 4
Green 2 2:00 -6:00 pm 4
4 Brooksville Stop bar 2 1 8:00 -12:00 pm 4
Green 2 8:00 -12:00 pm 4
5 Brooksville Stop bar 1 1 9:05 -11:45 pm 2:40
Green 2 9:35 -11:45 pm 2:10


are shown in Table 3-4.



Date

6 26 06
6 29 06
6 29 06
6 30 06
6 30 06
7 05 06
7 05 06
7 07 06
7 07 06
7 07 06
7 0706


i This indicates the field-of-view perspective of the cameras used for data collection at the site.









Data Reduction

The composite image videos were manually processed. The information that was recorded

from each video included the time when the signal turned green and the time when the front axle

of each vehicle in the queue crossed the stop bar. Also recorded was the type of each vehicle in

queue, according to a predetermined vehicle classification scheme (either passenger car, small

truck, medium truck, or large truck). Examples of the types of vehicles that fall into each of the

truck categories are shown in Appendix B The individual headway data and lost times were used

to develop the truck passenger car equivalent values for the three different classifications of

trucks, as discussed in Chapter 4.

Due to equipment problems, tapes 2 and 6 at the Williston Rd/34th St site unfortunately did

not have the start of green information. However, all other information could still be obtained

from these videos.

It should be noted that the precision of the timing measurements is limited to 0.0333

seconds. This is a result of using video cameras and recording equipment that utilize the

standard frame rate of 30 frames per second.

Simulation Modeling

This section describes the development of the simulation program, which was used for the

generation of a much larger data set to base the development of the PCE values upon. While

there are a variety of commercial simulation programs on the market that can simulate traffic

flow at signalized intersections, it was decided to develop a custom simulation program. This

decision was made for several reasons: 1) many commercial software programs do not readily

provide detailed documentation about the underlying car movement models; 2) the car

movement models contained in some programs are not very robust, particularly with respect to

heavy vehicles, or do not provide for user modification of some key parameters within those









models; and 3) with a custom simulation program, custom pre- and post-processing routines can

also be developed for increased efficiency. Thus, by developing a custom simulation program,

all aspects of the car movement models and other features of the program can be completely

controlled.

The development of the simulation program involved aspects such as creating a user

interface for specification of simulation scenarios and run settings and implementation of a car-

following model. These efforts are described in the following sections.

Car-Following Model

The foundation of a traffic simulation program is the underlying mathematical models that

describe the movement of vehicles along the roadway system. Thus, the first task in the

development of the simulation program was the selection of a car-movement model that was

suitable to a queue discharge situation at a signalized intersection.

Based upon the literature review in Chapter 2 for car-following models, and the needs of

this research proj ect, the Modified Pitt model was selected for implementation [1 l].

The Modified Pitt car-following model calculates an acceleration value for the trailing

vehicle based on intuitive parameters, such as the speed and acceleration of the lead vehicle, the

speed of the trailing vehicle, the relative position of the lead and trail vehicles, as well as a

desired headway. Car-following models are generally based on a 'driving rule', such as a

desired following distance or following headway. The Modified Pitt model is based on the rule

of a desired following headway.

As indicated before, the headway of each vehicle depends of the leading and trailing

vehicle and this model takes into consideration the physical and operational characteristics of

both. The main form of the model is shown again in Equation 3-2.











Kxrs, (t + R) sf (t + R) L, h x vf
+ v, (t +R) v (t + R) xT ~-
af (t + T) = 1[3-2]

,( 2xT1

Where :
a,~(t+T) = acceleration of follower vehicle at time t+T, in ft/s2
az(t+R)= acceleration of lead vehicle at time t+R, in ft/s2
sl(t+R) = position of lead vehicle at time t+R as measured from upstream, in ft
s,(t+R) = position of follower vehicle at time t+R as measured from upstream, in ft
vf(t+R) = speed of follower vehicle at time t+R, in ft/s
vl(t+R) = speed of lead vehicle at time t+R, in ft/s
L1 = length of lead vehicle plus a buffer based on jam density, in ft
h = time headway parameter (refers to headway between rear bumper plus a
buffer of lead vehicle to front bumper of follower), in seconds
T = simulation time-scan interval, in seconds
T = current simulation time step, in seconds
R = perception-reaction time, in seconds
K = sensitivity parameter (unit less)

For application to this proj ect, the value of the L parameter varied based on one of the four

different vehicle types. The time headway parameter (h) was set up as a random variable, rather

than a constant value, to introduce an additional stochastic element to the model. Its value was

based on a normal distribution to represent the more realistic scenario that desired headways vary

by driver. The mean and standard deviation for this distribution could be specified for each of

the four vehicle types. Thus, desired headways can vary by driver, as well as by vehicle

category.

Additional details on the Modified Pitt car-following model can be found in Cohen [24].

Program Development

The simulation program was written in Visual Basic 6. A screen capture of the main user

interface is shown in Figure 3-16.




















































Figure 3-16. Simulation program user interface

The top of the user interface is where all of the vehicle characteristics and model


parameters are specified. The lower part of the user interface is where the simulation run options


were specified. Four options can be specified:


1. Generate a queue of vehicles randomly, based on the specified vehicle proportions (the
queue length is a constant 8 vehicles).

2. Generate a specific number of each vehicle type in the queue, in random positions.


3. Generate a specific vehicle type in a specific queue position, for each of 8 total queue
positions.

4. Reading any number of pre-specified queue configurations from an input file.


-Vehicle. Driver. and Model Parameters-
Max
Veb Len Accel


Small Trucks ~
Medium Trucks ~
Large Trucks ~


Max FFS
D Iel Mean
I I (f/s]


Damping Factor [K) 12

Acclsbeera~tion [it/^2]
First Vehicle Reaction Time to Green

Mean Std. Dev.~


-Gueue Generation Scheme~
r Generate Vehicle Types Randomly


SGenerate~ehicle Type InSpecific
Queue Posit on


Run Options-


Passenger Cars
Small Trucks
Medium Trucks
Large Trucks


IPC ~
IPC ~




IPC ~


Num of Iterations per Cycle 11
Filename for output data
IExpDesign_out
p Output Summary Results Only
SShowy Simulation Animation

Seplec~#nt Cyclae # oAnmt
(lselct on)


Cycle # Iteration #


r Generate Specific Number of each
Vehicle Type [in Random Positions)
-Vehicle Input

Passenger Car 1
Small Truck
Medium Truck I
Large Truck
Total Queue Length


SRead Cycle and Queue Information
from File

IExpDesign csy

Browse...


~1117lnm*~u~tmm ~m;m ~-?rrl~~~ rLli-i~m;lrlmn;blnlY'~r;n~m;m:~w~mirrm r. I~IS11


FFS Std Hdwuy Hdwy R R Std Stop Gap Stop Gap
Dev Mean Std Dev Mean Dev. Mean Std Dev
(f/] (ec sc [e sd i] [t









The first three options were used primarily for model testing purposes. The fourth option

was used to facilitate the running of a large number of pre-specified simulation scenarios

according to the experimental design, described in a later section.

The program includes a traffic animation component. This component animates the

vehicle traj ectory information recorded during the simulation process. It updates the position,

speed, and acceleration of each vehicle in the queue every tenth of a second. This screen

includes the signal status and the elapsed time from the beginning of the green. Additionally, it

includes a table that displays key vehicle properties during each time step of the animation,

which is used primarily for diagnostic purposes. Screen captures of this animation screen are

shown in Appendix C.

Experimental Design

As mentioned previously, the main purpose of the Hield data was for calibration of the

simulation program. The simulation program was used to generate the data for the development

of the passenger car equivalent values for trucks, as well as revised lost time estimates. One

obvious limitation with the Hield data is that a limited number of vehicle type-queue position

combinations were observed, and of the combinations that were observed, some were observed a

very limited number of times, some only once. With the use of simulation, a wide variety queue

combinations (i.e., different vehicle types in different queue positions) can be generated, as well

as any number of replications of a specific queue combination.

One of the first decisions to make regarding the experimental design is which variables to

include. The study on saturation flow rate by Bonneson [8] identified several factors that affect

this value, such as speed limit, number of lanes, and traffic pressure. As discussed in Chapter 4,

none of these factors were found to be significant, or at least were inconclusive, in the field data.

Thus, none of these factors were included in the experimental design. Some of these variables









are still inputs to the simulation program, such as speed limit, but these were fixed at one

"average" value. Furthermore, given the anticipated size of the simulation analysis data set, it

was felt that simulation and analysis resources would be better spent focusing on just the truck-

specific aspects of the queue discharge mechanism, which was the primary concern for this

proj ect. With this approach, any revised PCE values resulting from this study could be used in

combination with the other factors developed as part of the Bonneson study. While the

Bonneson study did identify a single revised truck PCE value, the authors admit that the

examination of the effects of trucks on capacity was not a variable of primary concern; thus,

truck percentages were quite low in the Hield data they collected for their proj ect. In fact, one of

the specific recommendations from that study was to further investigate truck PCE values.

Therefore, the experimental design consisted of just varying the vehicle type by queue

position, for a fixed queue length, and leaving other factors Eixed at representative values. Note

that certain factors in the simulation program, such as maximum acceleration, are varied

randomly by vehicle/driver for each queue generation, according to a mean value and standard

deviation. However, these factors are part of the stochastic simulation process, and not factors to

be estimated as part of the analysis process.

In working within the framework of the HCM guidance for measuring lost time and

saturation flow rate, the queue length for the experimental design was set at eight vehicles for

each combination. Thus, with the four different vehicle types, and a queue length of eight

vehicles, the number of possible combinations is 65,536 (4 ). In order to reduce the

computational burden due to such a large number of combinations, and to better reflect reality

with the specific combinations, the number of combinations was reduced. Among the 65,536

possible combinations for four vehicle types and eight queue positions are many combinations









that include a high percentage of trucks in the queue. Since queues with a high percentage of

trucks were extremely rare in the Hield, it was decided to eliminate all queue combinations that

consisted of a truck percentage of more than 50%. Of all the Hield data queues, very few queues

had more than three trucks (out of a queue length of eight vehicles), and only one queue had Hyve

trucks. No queues had more than 5 trucks (out of eight). Elimination of all queue combinations

with more than four trucks resulted in a total of 7,459 combinations, which comprised the final

experimental design.









CHAPTER 4
ANALYSIS OF FIELD DATA AND CALIBRATION OF SIMULATION MODEL

This chapter describes the reduction and analysis of the field data, as well as the

development of a model to fit the traffic flow data obtained from the field. It also describes in

detail the calibration of the simulation program using the collected field data

Summary of Data Reduction

From the reduced field data, only signal cycles that had queues with at least 8 vehicles

were retained for data analysis. A summary of this data set is included in Appendix D. This

table shows the different queue compositions observed in the field along with their respective

frequencies. Note that for notational convenience, vehicle types are expressed in a numbered

format, where 1 is a passenger car, 2 is a small truck, 3 is a medium truck and 4 is a large truck.

This data set consisted of a total of 403 cycles, where 174 cycles were passenger cars only,

126 cycles had 1 truck, 68 cycles had 2 trucks, 28 cycles had 3 trucks and 7 cycles had 4 trucks

in queue. Some of these queue compositions per cycle were repeated; therefore there were a

total of 110 different observations. In addition, only one queue was observed to have 5 trucks

but it was left out of the calibration process since the experimental design was constrained to no

more than 4 trucks in queue. This will be explained in more detail later in this chapter.

In chapter 3 it was discussed that the time headway value between successive vehicles is a

function of both the leading vehicle and the trailing vehicle. A summary of the average time

headway for the 16 possible vehicle pairs (for 4 vehicle types) is shown in Appendix E. In Table

4-1 is a summary of those tables for vehicles in queue positions 2-8 and 5-8. Headways of

vehicles in position 1 are not included in these tables since only 23 1 out of 403 cycles included

the start of green.











Table 4-1. Average headway and frequencies for each leader-follower combination
Average headways of vehicle pairs in positions 2 through 8
Trail
vehicle' 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4
Lead
vehicle' 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
Frequency 417 51 37 61 45 8 5 7 36 4 4 1 65 6 7 16
Mean 2.40 3.13 3.34 4.70 3.01 3.85 4.92 5.61 3.67 4.86 5.70 4.24 4.14 4.42 4.97 5.09


Average headways of vehicle pairs in positions 5 through 8
Trail
vehicle' 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4
Lead
vehicle' 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
Frequency 232 29 21 38 28 3 4 3 21 2 3 1 39 2 3 11
Mean 2.19 2.82 2.72 4.13 2.86 3.08 4.88 4.22 3.74 5.61 4.59 4.24 4.13 4.50 4.46 5.23
1 .
Vehicle Types: 1 = passenger car, 2 = small truck, 3 = medium truck, 4 = large truck

Time Headways

Time headways are often used to estimate the impact of trucks in the traffic stream, but to


specifically calculate the saturation flow rate it is necessary to use the average headway of


vehicles in positions 5 through 8 (hs_,) or also known as the saturation headway (hSAT). It should


be noted that this definition is consistent with the HCM, although some studies have found that

the saturation headway is not achieved until the sixth or later vehicle in the queue. The concept

of the saturation headway, consistent with the HCM definition, is illustrated mathematically with

the following equations.

1. Saturation headway (i.e., the average headway for vehicles in positions 5-8)

hS = [4-1]

2. Average headway for vehicles in positions 1-8
T,
hz~s= [4-2]

3. Average headway for vehicles in positions 2-8

h2-8_ [4-3]

Where :
T, = the time it takes for vehicle i to cross the stop bar

Table 4-2 summarizes the headways calculated from the field data.










Table 4-2. Time headways from field data
All vehicles All vehicles with start-up time Passenger cars only Passenger cars only with start-up time
hsar 2.87 2.83 2.18 2.23
hi-8 N/A 2.98 N/A 2.48
hZ-s 3.11 N/A 2.36 N/A


Several models were specified in the statistical software package STATISTICA [25] to

attempt to reasonably describe the impacts of trucks on queue discharge for these field data.

These models considered queue composition, truck percentages, and site characteristics. The

specifications were done using a non-linear regression analysis with a confidence level of 95%.

A summary of this model development is described in the following paragraphs.

Model 1

The first model (Equation 4-4) used an alternate form of the fil equation (HCM chapter

16) to estimate the average saturated headway (hSAT). This analysis used the data that did not

contain start-up reaction time values.


h,= hSPCx (1+Pctsr x b, -1)+ Pctz, x(b, -1)+ PctLT x (b -1)) [4-4]

Where :
hSAT = saturation headway
hSPC = Saturation headway of passenger cars in a passenger car only queue
Pctsr = percentage of small trcks inqueue
Pctnnr = percentage of medium trucks in queue
PctLT = percentage of large trucks in queue

In this model, the coefficients bl, b2 and b3 TepreSent the ET ValUeS foT Small, medium and

large trucks respectively. The estimated coefficient values are shown in Table 4-3. An R2 ValUe

of 0.3 5 91 was obtained for thi s model.

Table 4-3. Results from STATISTICA for model 1
Coefficients Estimate Standard error t-value
hSPc 2.216044 0.033277 66.59442
bi 1.655451 0.165439 10.00640
b, 1.771816 0.196516 9.01615
b3 2.677482 0.132386 20.22480











Model 2

The second model (Equation 4-5) is the same as Model 1, but this model estimates the

average headway for vehicles in position 2 through 8, where hPC is the average headway for


passenger cars. Since some of the data did not include the start of green time (thus the headway

for vehicle 1 could not be determined), this analysis was based on a larger data set. The

estimated coefficient values are shown in Table 4-4. An R2 of 0.5923 was obtained for this

model .


h2-8= hPC x 1+Pct, x (b, -1)+ Pcts, x (b2 -1)+ Pctx (b3 -1)) [4-5]

Table 4-4. Results from STATISTICA for model 2
Coefficients Estimate Standard error t-value

hPc 2.381300 0.021310 111.7433

bl 1.765421 0.099050 17.8236
b2 2.062019 0.118059 17.4660

b3 2.508828 0.077991 32.1682


Model 3

The third model (Equation 4-6) has the same characteristics as M~odel 1 but using only the

data that contain the start-up time. The estimated coefficient values are shown in Table 4-5. An

R2 Of 0.3798 was obtained for this model.


hSAT = hscx (1+ Pctsr x (b, -1)+ PctM, x (b2 -1)+ Pct,, x (b3 -1)) [4-6]

Table 4-5. Results from STATISTICA for model 3
Coefficients Estimate Standard error t-value

hSPc 2.204078 0.040422 54.52684
bi 1.705201 0.189231 9.01121
b2 2.148228 0.217222 9.88953
b3 2.639080 0.174277 15.14304










Model 4

The fourth model (Equation 4-7) has the same characteristics as M~odel 2 but using only the

data that contain the start-up time. The estimated coefficient values are shown in Table 4-6. An

R2 of 0.5309 was obtained for this model.


~,= h,,c x (1 + PctST x (b, 1)+ PctMT x (b, 1)+ PctL Tx (b, 1)) [4-7]

Table 4-6. Results from STATISTICA for model 4
Coefficients Estimate Standard error t-value

hPC 2.479479 0.024889 99.61976
bl 1.414669 0.102169 13.84634
b7 1.882030 0.117720 15.98733
b3 2.239560 0.093155 24.04133

These models resulted to have a low value for R2 and it was due to the variance inherent in

the data. This could be due to insufficient field data. In addition, there were specified several

other models that included additional variables such posted speed limit, area type, and number of

lanes. These models were not found to be statistically significant.

Start-up Reaction Time (SRT)

The Start-up Reaction Time (SRT) is the time from start of green until the first vehicle

begins to move. This measurement is sensitive to the timing precision of the video frame rate

(0.0333 frames/sec). If vehicles in the first position had their front axle on the stop bar, their

headways were set equal to the SRT. Cycles that had vehicles in the first position with their

front axle beyond the stop bar were dropped out of the data set.

These times were obtained from the videos and were measured only from site 1. Site 1 has

two through lanes, but these data were collected during a time when one through lane was

closed. After careful observation and comparison between the queue discharge at this site for

two through lanes and one through lane, it was concluded that this particular lane closure set-up













did not have any adverse impact on start-up reaction times or queue discharge rates. Right


turning vehicles were not included in these measurements.


A frequency graph of the field-measured SRT values measured is shown in Figure 4-1.



35-

30-

25-

~20-

u-15 -

10 --




125 175 225 275 325 375 425 475 525 575 625 675
Seconds


Figure 4-1. Start-up reaction time frequencies


Based on the observations, it was decided that a maximum of 3.5 seconds can be


considered as a realistic maximum start-up reaction time. SRTs greater than 3.5 seconds were


typically due to an obvious hesitation (related to distraction or similar) of the drivers. Therefore,


SRT values above 3.5 seconds were trimmed from the data set. Figure 4-2 shows a frequency


distribution for only those drivers with a SRT of 3.5 seconds or less.



35-

30-

25-

20-









125 15 175 2 225 25 275 3 325 35
Seconds


Figure 4-2. Start-up reaction time frequencies for trimmed data set











Using this trimmed data set with 134 observations, a mean start-up reaction time of 2.04

seconds and a standard deviation of 0.47 were calculated. This mean value was somewhat

consistent with previous studies reported in the literature. Kockelman [6] reported a mean SRT

value of 1.79 sec and Li and Prevedouros [10] reported a mean SRT value of 1.76 and standard

deviation of 0.61. They also concluded that these values were normally distributed.

For simulation purposes, it was assumed that the first vehicle started at the stop bar and

therefore the headway of the first vehicle in queue was equal to the SRT regardless of vehicle

type. From field data and other studies it was decided to the set SRT equal 2.0.

Start-up Lost Time (SLT)

The Start-up lost time is the difference between elapsed time for first four vehicles (1-4) to

cross the stop bar and the time for the last four vehicles (5-8) in queue assuming passenger car

only queue. The SLT can be calculated as it is shown in Equation 4-8.

SLT = TT, (h, x 4)
1-4 S,4T[4-8]

Where :
TT1-4 = total time required for first four vehicles in queue to cross the stop bar
h,, = average saturation headway for a passenger cars only queue

The field data were examined to obtain some sense of whether or not trucks are more or

less likely to be in the first four positions of the queue. Table 4-7 summarizes these data.

Table 4-7. Distribution of trucks by site and position in queue
Site indicator
1 2 3 4 5 6-1 6-2 6-3 6-4 Total
Total cycles 50 11 24 17 22 52 66 18 143 403
Cveles with tricks 41 11 8 14 20 20 25 9 81 229
Cycles with tricks up-front* 6 1 3 8 11 9 10 7 40 95
00 of tricks up-front* 1500 900 3800 57oo 5500 4500 4000 7800 4900 4100
up-front = in the first four positions










Based on the Hield collected data, it was not possible to draw a definitive conclusion about

whether trucks are more or less likely to be at the front of the queue. Thus, for the purposes of

this study, it was assumed that trucks were randomly distributed throughout the entire queue, for

each cycle. Additionally, it was observed that the Start-up Lost Time increased proportionally

with an increase in overall truck percentage.

Calibration of Simulation Model

As mentioned before, for simulation to be effective, it is necessary to calibrate a simulation

program against Hield data. An extensive calibration process was performed in order to identify

the parameter values that resulted in the simulation data providing the best match with the Hield

data. The quantitative measure used in this process was the Mean Square Error (MSE). The

MSE is commonly used to compare a predicted value with an observed value. In this case, it was

used to compare the observed average headway from the Hield data with the average headway

estimated from the simulation data. Equation 4-9 shows the mathematical form for MSE.



M~SE = sit [4-9]

Where :
M~SE = Mean Square Error
hF~eld = average headway from the Hield data
hsmm; = average headway from the simulation data
N = total number of observations

The calibration process consisted of comparing valid field data cycles with simulation

cycles with the same configuration. Valid cycles were those cycles with at least 8 vehicles in

queue and that showed a normal behavior. Cycles with implicit hesitation of the drivers, lane

changing, first vehicle beyond the stop bar, motorcycles and right turns were considered as not

valid or non-normal behavior.










In this process there were a couple of parameters that remained fixed. The simulation

time-scan interval (T) is not a direct calibration parameter and was not changed throughout the

process. Also parameters as R and K were fixed at 0.7 and 1.25 respectively. These values were

recommended by Cohen [1l]. The vehicles' lengths were fixed using values that were based on

observations from this study's field data, and were also consistent with values recommended by

Long [22]. The rest of the parameters were adjusted during the process. A screen capture of the

program with the final set of parameters is shown in Figure 4-3.

Vehicle. Driver. and Model Prmtr
Max Max FFS FFS Std. Hdwy Hdwuy R R Std. Stop Gap Stop Gap
Veh Len Accel. Decel. Mean Dev. Mean Std.Dev. Mean Dev. Mean Std Dev. Damping Factor (K) 1.25

Small Trucks 1 01 5 1 1 7 1 35 5 5 71 4 1 2 Acceleration [it/s^2]
Medium Trucks Z i FirSt Vehicle Reaction Time to Green
Large Trucks 1 51 3 1 31 5. .5 2 71 01 25 Mean Z Std Dev

Figure 4-3. User-adjustable vehicle, driver, and model parameters

In the process of calibration and trying to achieve a minimum MSE value, it was necessary

to keep in mind that the final values of the parameters should stay within a reasonable range. For

example, if a minimum MSE value is obtained by using a passenger car free flow speed of 30

ft/sec (~20 mi/h), it is not realistic to say that a passenger car will cruise at such a low speed.

Process of calibration: An input file was created and read into the program to facilitate

the process. This file contained each queue composition observed in the field. A first attempt

included ten replication runs of each queue scenario. The variance within the runs was too high

and the required sample size was on the order of 60 replications, for that reason the number of

replications was increased to 100.

During the process of calibration the parameters that were primarily experimented with

were the lead vehicle's acceleration and the maximum acceleration, free-flow speed, minimum










desired headway and inter-vehicle spacing per vehicle type. A total of 99 different parameter

combinations were run in an effort to obtain the lowest MSE. In those iterations, the value of the

MSE fluctuated from 0.097 to 0.415 for the average headway of vehicles 2 through 8 and from

0. 142 to 0.761 for the average saturation headway. The final set of parameters had 0. 118 and

0. 142 for the average headway of vehicles 2 through 8 and the average saturation headway

respectively. The final set of parameters is shown in Table 4-8.

Table 4-8. Final choice of calibration parameter values
Headway Stop gap
Vehicle length Maximum acceleration FFS FFS std. dev. Headway std. dev. Stop gap std. dev.
(ft) (ft/S2) (ft/s) (ft/s) (sec) (sec) (ft) (ft)
PC 15 10 72.50 3.75 1.50 0.25 10 2.0
ST 30 5 67.50 3.75 2.50 0.25 14 2.0
MT 45 4 62.50 3.75 3.00 0.25 16 2.5
LT 65 3 57.50 3.75 3.50 0.25 20 2.5

Execution of Experimental Design

Once the experimental design was complete, an input file was generated with the 7459

queue combinations. The simulation program was run with this file performing 100 replications

of each queue combination. Based on some testing, it was decided to perform 100 replications to

account for the considerable variance inherent in this process. Since there was considerable

variance in the field data, the calibration process of the simulation still allows for a large amount

of variance in each individual simulation run. The choice of 100 replications was somewhat

conservative, but not extremely conservative.

An average of each of the 100 replications was calculated and used in the analysis

database. This new data set was inputted in the statistical software package STATISTICA.

Subsequently, variables were created as needed to run the models. These variables include, but

are not limited to, indicator variables for each vehicle type per position, indicator variables for

vehicle pairs per position, frequencies and percentages of each one of theses scenarios.









Additional experiments were run to verify certain numerical results, such as the following.

* 10,000 passenger car-only queues to verify the base saturation headway

* 400 passenger car-only queues, changing the vehicle in position 4 to estimate the effect of
different truck types on the headway of the vehicle in position 5.

* 256 cases with 100 replications of trucks in positions 1-4 and PCs in positions 5-8 to
isolate the effect of trucks on the start-up lost time section of the queue.

* A data set with fabricated fixed values for each vehicle pair, regardless of their queue
position, to verify model formulations.









CHAPTER 5
ANALYSIS OF SIMULATION DATA

This chapter describes how the data generated from the simulation experimental design

process were used to obtain the new PCE factors. For the purposes of this research, the queue

was analyzed in sections that were consistent with the guidance given in the HCM 2000 with

regard to the components of start-up lost time and saturation headway. That is, the first four

headways are included in the start-up lost time, and the headways of the remaining vehicles

represent the saturation headway. Thus, the queue was subdivided into two sections for purposes

of the analysis. Figure 5-1 illustrates these two queue sections.






Saturated Section of Queue Start-Up Lost Time Section


Figure 5-1. Queue sections for analysis

In reality, the lost time does not necessarily end with the headway of the fourth vehicle, but

for purposes of this research it was decided to be consistent with the HCM framework.

Furthermore, the field data collected in this study were not conclusive with regard to this issue.

As previously mentioned, the analyses were done assuming trucks were distributed

randomly within each queue. That is, for any given cycle, a given number of trucks may appear

at any location throughout the queue. In some queues, a given number of trucks may be at the

front of the queue, in some queues this same number may be in the middle of the queue, and so

on. But over the length of the analysis period, for a fixed percentage of trucks in the traffic

stream, it is assumed that any position within the queue is as likely to have a truck in that

position as any other position in the queue.









The first approach to find the PCE values was to analyze the saturated part of the queue, as

it is recommended in the HCM. A simple model using only the percentages of each type of truck

in positions 5 through 8 was specified. The results of this model did not reflect exactly the

behavior of the queue. This was expected since this model does not explicitly account for the

headways being a function of the specific leader-follower combination. Subsequently, another

model was specified using the frequency of each vehicle pair in the second half of the queue.

This approach assumed that each vehicle pair had the same headway regardless of their position

in the queue. The estimates from this model did not replicate the performance of the queue as

closely as hoped. To confirm that the model was specified correctly, a different program was

developed to generate fixed headway values that were independent of queue position. When the

same model was run on this data set, it provided a better fit, but still not exact. Thus the next

step was to incorporate indicator variables for each one of the 16 different vehicle pair

combinations, for positions 5 through 8. This model demonstrated that each vehicle pair has a

different headway value depending of their position in queue.

It was also noted that the type of vehicle in position 4 had a significant impact on the

calculation of saturation headway, as expected. Furthermore, it was also observed that trucks

present in the first few positions of the queue also influence the saturation headway. Figure 5-2

shows a diagram with three different scenarios reflecting these impacts.







Figure 5-2. Impact of trucks in the first part of the queue to the saturation headway

The first case (on top) has only passenger cars, the second case (middle) has a large truck

in position 4, and the third case (bottom) has large trucks in positions 3 and 4. Although the









second part of the queue was the same for all the cases (passenger cars in positions 5-8), the

saturation headways were significantly different. The saturation headways for these three

scenarios were 2.033, 2.795 and 2.641 seconds respectively. The only difference between the

first and second cases is the presence of a large truck in the fourth position. The size of this

vehicle impacts directly the headway of the vehicle in position 5. In addition, large trucks have a

lower acceleration rate when compared to passenger cars, which affect the acceleration of the

vehicles behind. In this case, vehicles in positions 5 through 8 were directly impacted by the

truck. These vehicles were constrained to the acceleration of the large truck since lane changing

was not allowed. This comparison confirms that the saturation headway is affected by the

vehicle in position 4.

Moreover, it was observed how the impact on saturation headway due to the number of

trucks in the first few queue positions might be overestimated. The impact on saturation

headway is just as much a function of the position of the trucks in the queue as it is the total

number of trucks in the first four positions of the queue. These data showed that the queue with

two large trucks had a smaller saturation headway than the one with only one truck. This

confirms that not only the vehicle type in position 4 impacts the saturation headway, but also the

presence of other trucks preceding the truck in position 4 provide an additional impact. In this

example, the difference results from the length of the trucks rather than their acceleration

capabilities. Vehicles in positions 5 through 8 were constrained to the acceleration rate of the

large trucks, but in the third case the vehicles were farther back in the queue. The distance to the

stop bar of each vehicle was larger and with the same acceleration the vehicles reached higher

speeds prior to crossing the stop bar. The spacing was very similar between vehicles in the two

cases, so with a higher speed and similar spacing, the result is lower headways. This result is









consistent with some studies that have claimed that the headways between vehicles at the back of

a long queue will be lower than those near the front because the vehicles near the back have

more time accelerate to a higher speed. However, the Li and Prevedouros [10] study found that

this is not always the case because headway expansion may occur toward the end of long queues

In light of these findings, it was desired to test a model that included indicator variables for

each one of the 16 different vehicle-pair combinations for each queue position. However, this

results in a model with an extremely large number of parameters, which was beyond the

capabilities of the statistical software package. Therefore, it was decided to treat each part of the

queue separately. A new data set was created with only passenger cars in the second part of the

queue (positions 5-8). This data set included 256 (44) different queue combinations, as it was

only the first four positions of the queue that varied by vehicle type. This new data set was used

to estimate the impact of trucks to the SLT.

The more accurate method to analyze the impact of trucks to the SLT is to use indicator

variables for the 16 different vehicle-pair combinations in each one of the first four positions.

But again, this approach was not practical. Therefore, since the simulation program provides the

same headway value for the first vehicle regardless of what type it is, it was decided to run a

narrower model with only indicator variables (for the 16 combinations) for the first two vehicles.

This model isolates the impact of the first vehicle in queue. The headway of the first vehicle was

the same for each vehicle type, but the effect to the trailing vehicle due to the length and

acceleration of the first vehicle varied. After the results from this model were obtained, this

model was included in another model that included indicator variables of vehicle types for in

positions 2 through 4. This model is shown in Equation 5-1.


T7-4 -27; TPC3-4 + C (bk x S'T + bk x M~T + bk x LT) [5-1]









Where :
TT1-4 = clearance time for the first four vehicles in queue
TT1-2 = clearance time for the first two vehicles in queue
TTPC3-4 = clearance time for passenger cars in positions 3 and 4
bik = additional time for vehicle type i in position k to clear
ST = indicator variable for vehicle type small truck (1-yes, 0-no)
M~T = indicator variable for vehicle type medium truck (1-yes, 0-no)
LT = indicator variable for vehicle type large truck (1-yes, 0-no)
k = indicator for queue position
i = ST for small truck, MT for medium truck, LT for large truck

And the clearance time for the first two vehicles was defined as is shown in Equation 5-2.


TT1-2 =2.0 +hPC-PC + [ b, x I,) [5-2]

Where :
TT1-2 = clearance time for the first two vehicles in queue
hPC-PC = headway of a passenger car in position 2 following a passenger car
b, = additional time for vehicle i following vehicle type j to clear
IJ = indicator for vehicle type i following vehicle type j
i = indicator for vehicle type
j = indicator for vehicle type

The final form of the equation to estimate the total time to clear the first four vehicles in

queue (Equation 5-1) is shown in Equation 5-3.
1.62 x b21 + 3.00 x b21 + 5.04 x b21

+1.38 x b221 + 3.09 x b222 + 4.53 x b223 1.47 x b32 + 2.49 x b33
T7;-4 = 10.59 + + 6.65 x b224 + 2.47 x b231 + 4. 19 x b232 + + 4.03 x b34 + 0.83 x b42 [53
+ 5.65 x b 233 + 7.80 x b 234 + 4.22 x b 241 + 1.21 x b43 + 1.72 x b, 44-3
+59x242 +73x243+94x244

Where the total time to clear the first four vehicles, if are passenger cars, is 10.59 seconds

and the bkz, and b, follow the format described earlier in this chapter.

From Equation 4-8, the SLT can be calculated as the difference of this time to clear the

first four vehicles and the equivalent of four times the saturation headway of passenger cars. The

SLT calculated for passenger cars only queue is equal to 2.47 seconds, which is higher than the









2.0 seconds recommended in the HCM 2000 [1], but somewhat consistent with the findings of

the field data.

It is also important to account for the impact of trucks to the SLT since the presence of

these vehicles increased significantly the SLT. In example, if the composition of the queue is as

it is shown in Figure 5-3.







Figure 5-3. SLT example

where a large truck is in the second position, a medium truck in the fourth position, and the

rest are passenger cars, the SLT can be estimated as indicated in Equation 5-4.

SLT = [10.59 + (4.22 x b241 )+ (1.72 x b44)]1- (4 x 2.03) [5-4]

Then, the SLT for this case of mixed traffic is 8.41 seconds. This value is considerably

higher when compared to the 2.47 seconds for passenger cars only queue and it needs to be

treated separately from the saturation headway.

Since it is not realistic to expect the practitioner to collect every vehicle type in queue and

their exact position, an equation was developed to estimate the SLT based on the percentage of

trucks in the queue. As discussed earlier, it is assumed that the trucks are distributed randomly

throughout the queue. Therefore, over many cycles with trucks distributed randomly within the

queue for each cycle, it can be assumed that on average, trucks will be distributed evenly

throughout the queue. The SLT equation is given by,

SLT = 2.5 +5.0x Pcts x 9.0x Pctz, +15.0x PctLT [5-5]

Where :
Pct = percentage of truck type i











After analyzing the two different sections of queue, a compound model was developed.


This model was used to obtain an estimate of the total time elapsed for clearing a queue length of


eight vehicles, using the experimental design data set. A general form of this model is shown in


Equation 5-6.


TT,_g = TT1-4 + 4 x hPC-P + (bh x F ) [5-6]
Where :
TT;_B = clearance time for the first eight vehicles in queue
TT1-4 = clearance time for the first four vehicles in queue
hPC-PC = headway of passenger car following a passenger car
b, = additional headway for vehicle pair (vehicle i following vehicle j)
F, = frequency of vehicle pair (vehicle i following vehicle j) in positions 5
through 8


The model estimation output is shown in Table 5-1 where bo is hPC-PC and the remaining


betas follow the "b," format described before.


Table 5-1. Model estimation results for additional headways of 16 vehicle pair combinations
Coefficients Estimate Std. error t-value

hPC-Pc 2.028586 0.004183 484.9803

bl2 0.590274 0.012212 48.3346

bl3 1.046166 0.012212 85.6654

bl4 1.852489 0.012212 151.6913

b21 1.024162 0.012212 83.8636

b22 1.531207 0.016902 90.5956

b23 2.035979 0.020033 101.6312

b24 2.993811 0.020033 149.4440

b31 1.407661 0.012212 115.2665

b32 1.920917 0.020033 95.8877

b33 2.393117 0.016902 141.5915

b34 3.377253 0.020033 168.5845

bg; 1.823759 0.012212 149.3388

b42 2.426886 0.020033 121.1444

b43 2.835042 0.020033 141.5186

644 3.573751 0.016902 211.4451










These estimates were substituted in Equation 5-6 and the final form of the model is defined

in Equation 5-7.


0.590 x F,, +1.046 x F,, +1.852 x F14 +1.024 x F,,

T7; = 7;,+ x .02 ++1.531 x F, + 2.036 x F, + 2.994 x F,4 +1.407 x F,,4

+ 2.427 x F42 + 2.835 x F43 + 3.574 x F44


Figure 5-4 shows a plot of this equation against the observed simulation data.


[5-7]


18 20 22 24 26 28
Observed


30 32 34 36


Figure 5-4. Total time for vehicles 1-8 using vehicle pairs in positions 5-8

This model provides the headway for a passenger car following a passenger car (hPC-PC) in

positions 5 through 8 and also the additional time headway taken by any other vehicle pair in

these positions. Therefore, it was possible to calculate the headway for each one of these pairs.

Table 5-2 shows these headways. In this table the PCE value for each vehicle pair is also shown.

These PCE values were calculated as a relative headway to the hPC-PC per the definition in the

HCM 2000.
















































































Figure 5-5. Time consumed by a large truck


i;li

iiiiiiiilililililililililililililililili
LMe4
iiiiiiiillllllllllll11111111111111111111
hr


Table 5-2. Headways and PCE values for 16 vehicle pair combinations
Vehicle pair hPC-PC Additional headway* Headway PEvle fec 6vhcepi
training leading (seconds) (seconds) (seconds)
PC PC 2.029 2.029 1.000
PC ST 0.590 2.619 1.291
PC MT 1.046 3.075 1.516
PC LT 1.852 3.881 1.913
ST PC 1.024 3.053 1.505
ST ST 1.531 3.560 1.755
ST MT 2.036 4.065 2.004
ST LT 2.994 5.022 2.476
MT PC 1.408 3.436 1.694
MT ST 1.921 3.950 1.947
MT MT 2.393 4.422 2.180
MT LT 3.377 5.406 2.665
LT PC 1.824 3.852 1.899
LT ST 2.427 4.455 2.196
LT MT 2.835 4. 864 2.398
LT LT 3.574 5.602 2.762

*Ah = (Veh, Veh,)-(PC PC)


Although these PCE values reflected more accurately the behavior of the queue, it is not


practical to have a PCE value for each possible vehicle-pair combination. In an effort to simplify


this model, a new set of PCE factors was derived for each one of the truck vehicle types. This


new set takes into consideration their headway and the time each vehicle type adds to the


following vehicle's headway. This concept can be described as the time consumedby any


vehicle type in queue. Figure 5-5 illustrates this concept with an example of a Large Truck in


queue. This example compares the headway of a PC following a LT with the headway of a PC


following a PC.










As shown, a LT adds an additional headway (Ah) to the PC. In this case the additional

time can be defined as it is shown in Equation 5-8.

AhLT = hPC->LT hPC->PC [ 5-8 ]

From the values in Table 5-2, 1.852 seconds is added to the headway of the passenger car

when following a large truck. This additional time should be considered part of the LT headway

rather than the PC since this additional time it is not present in a PC only queue. Also shown is

the time consumed by a LT in queue (dark shade) which was defined as the time headway of a

LT plus the additional headway this vehicle type adds to the following vehicle.

The Ah of each vehicle type is not only added to the PC. These impacts are present

regardless of what type of vehicle is following. Table 5-3 includes a summary of each vehicle

type's headway when following a PC and the Ah they add to any other vehicle type.

Table 5-3. Vehicle type's headway and their Ah
Leading
Trailing Headway when following PC ah ST ah MT ah LT
PC 2.029 0.590 1.046 1.853
ST 3.053 0.507 1.012 1.970
MT 3.436 0.513 0.986 1.970
LT 3.852 0.603 1.011 1.750
Average ah 0.553 1.014 1.885

In order to develop a model that does not take into consideration the position of each

vehicle in queue, it was necessary to use the average Ah per vehicle type. By using the average

of the Ah, the proportion of these vehicle pairs in queue are not being taken into consideration.

To develop PCE values based on just three truck categories, rather than all 16 possible lead-

follow vehicle pairings, the resulting PCE values will obviously be generalized. That is, with

this approach, information about specific position in queue for each truck is not utilized.

However, the reason, again, for this approach is that it is expected that practitioners will not

collect data at this level of detail. With such a generalized approach, this loss of information










requires some level of approximation. The implicit assumption is that each vehicle pair occurs

with the same frequency in the traffic stream. While this may rarely be the case, the obj ective

was to arrive at generalized PCE values that yielded an fHV value that still tracked reasonably

well with the fHV value that results from using all 16 PCE values, for a varying range of overall

truck percentage, as well as varying relative truck type percentages. This more general equation

using the time consumed (H) by any vehicle i in queue is defined in Equation 5-9. Where vehicle

j is following vehicle i.

HI = ht->PC + Ah~ [5-9]
Where :


Ah = 4''= [5-10]

Andl~= PC, 2 =ST, 3 =MTand 4 =LT

Shown in Table 5-3 are the values for the time consumed in queue by each vehicle type.

These values can be used to arrive at new PCE factors since this time consumed by each vehicle

type can be considered as their time headway. Table 5-4 shows the PCE factors that were

calculated as the relative time consumed per each vehicle type in queue when compared to a PC.

Table 5-4. Time consumed and PCE values for each vehicle type
Vehicle type h~,-c (seconds) Avg. ah, (seconds) H, (seconds) PCE,
PC 2.029 0.000 2.029 1.000
ST 3.053 0.553 3.606 1.778
MT 3.436 1.014 4.450 2.194
LT 3.852 1.885 5.738 2.828


To assess the accuracy of these values, a model of total time needed to clear a queue with

eight vehicles with only three types of trucks, instead of all vehicle-pair combinations, was

tested. The model was specified in STATISTICA and the general form of this new model is

shown in Equation 5-11.

T7;, = TE-4 + 4 x hPC5-8 + (bsr x Fs,, + b,, x F, + b,, x F, ) [5-11]











Where :
TT,_B = time needed to clear the first eight vehicles in queue
TT1-4 = time needed to clear the first four vehicles in queue
hpcs-s= headway of passenger cars following passenger cars in positions 5
through 8
b, = additional headway for vehicle type i
F, = frequency of truck type i in positions 5 through 8
i = ST for small trucks, M~T for medium trucks, and LT for large trucks

The model estimation results are shown in Table 5-5, where bo represents the time of only

PCs in positions 5 through 8 and the remaining betas follow the "b," format described before.

Table 5-5. Results from STATISTICA for model with vehicle types
Coefficients Estimate Std. error t-value
hpcs-s 2.270076 0.029382 309.0408
bsr 1.228277 0.018823 65.2534
bhrr 1.952747 0.018823 103.7416
bLT 2.985894 0.018823 158.6285


Figure 5-6 shows the plot of this equation against the observed data.





36 *

34 *

32 *.

-cr 30









20

10
18 20 22 24 26 28 30 32 34 36
Observed



Figure 5-6. Total time for vehicles 1-8 using vehicle types in positions 5-8










This figure illustrates the increase in variance when compared to the plot based on all 16

vehicle pairs, as expected. Using just the three general categories of truck type instead of all 16

vehicle pairs obviously results in less precision of the saturation flow rate estimate. Table 5-6

shows the resulting headways and PCE values calculated from the results shown in Table 5-5.

Table 5-6. PCE factors for three vehicle types
Vehicle type headway (seconds) PCE
PC 2.270 1.000
ST 3.498 1.541
liT 4.223 1.860
LT 5.256 2.315


Note that these PCE values are considerably lower than the ones calculated from the other

model shown in Table 5-4. These values demonstrate that the impact of the passenger cars is

being overestimated and the impact of the trucks is being underestimated. As previously

discussed, a truck in position four has a significant impact to the saturation headway, and this

approach did not take that into consideration.

Another alternative was to use the alternate form of the fHv equation from the HCM. The

following derivation was made with the purpose of relating the current analysis to the HCM' s

approach. Starting with the model specified to estimate the total time needed to clear the first

eight vehicles in queue Equation 5-11, and if TT,_, = TT,_4 + TT,, then,


TT_, = 4 x hPC-8 + (b, x F, + b,,, x F,, + bLT FLT) [5-12]

Again, the HCM defines the PCE as a relative headway. Since hees-s is the headway of a

passenger car and b, is the additional headway of vehicle type i, then



PCE, =I PC-8 -> b = (PIE, x hPC5-8)-hPC5-8 [5-13]
hPC5-8









Then,

TT,=4xpc'(hecs-s x PCEsr heess) x Fsr +
TE_ = ee- (hecs-s x PCE,,,- heess) x F,,+j [5-14]
(hecss x PCELT hees-)x FLT

TT
hey, = 6- [5-15]
and

Pct, =~ [5-16]

then simplifying the equation and dividing by 4 to obtain one average headway value

yields,

hSi7 = hP15-8 x [1 + Pcts. x (PC~sr -1)+ Pctz,, x (PCE,,, -1)+ PctLTl x (PCELT -1)1 [5-17]

This model was used to generate the PCE values. The results were similar to those

obtained when using vehicle types in the total time formulation (Equation 5-11). This model was

also run with a Eixed value for hPC-PC and the results were similar to the ones obtained when

using vehicle pairs (Equation 5-7). From these results it was concluded that the impact of trucks

in the first few positions of the queue was not accounted for. In addition, the data set was

expanded adding passenger car only queues to reduce the overall truck percentage from 46% to

10%. Then the same model was run (Equation 5-11) and the results were the same as fixing the

hPC-PC. IHCreaSing the amount of passenger car only queues in the data set had a significant

change to the estimation of the saturation headway.

The alternate form of the equation forfer can also be derived from the model based on all

16 vehicle pairs. Equation 5-18 gives the Einal form of this equation.



h5-8 = hPC-PC (7Zki~= x PE, 1i) [5-18]











Model estimation results are shown in Table 5-7, where bo estimates the hPC-PC and the


remaining betas estimate the PCE for vehicle i following j.


Table 5-7. Model estimation results for fHv equation with 16 vehicle pairs
Coefficients Estimate Std. error t-value
hPC-PC 2.099136 0.002592 809.6979
bl2 1.255759 0.003755 334.4511
bl3 1.474404 0.003898 378.2679
bl4 1.860251 0.004181 444.9291
b21 1.431489 0.003869 370.0254
b22 1.672034 0.005277 316.8362
b23 1.910199 0.006270 304.6783
b24 2.364296 0.006512 363.0607
b31 1.614128 0.003996 403.9260
b32 1.855496 0.006243 297.2072
b33 2.079154 0.005505 377.6886
b34 2.545657 0.006620 384.5518
b41 1.813318 0.004145 437.5067
b42 2.097403 0.006365 329.5371
b43 2.290151 0.006470 353.9746
b44 2.638246 0.005874 449.1694


Although this form is consistent with the HCM, to base the new PCE values it was decided


to use the model (Equation 5-7) that estimates the total time for the first eight vehicles in queue


to clear with the 16 different vehicle pairs. Accounting for the total time provides a better


understanding of what occurred in the entire queue and it isolated the impact of trucks to the SLT


from the saturation headway. However, the main objective of this research was to develop new


PCEs for three truck categories that result in an fir7 value that match reasonably well with the


values based on all 16 PCEs. It is not possible to have only one set of PCE values (for just the


three truck categories) and have them result in the same~ fzvalue as for the 16 PCE values across


a range of truck percentages.


In order to obtain the new PCE values for the three trucks types from the sixteen vehicle


pairs' headway, two different methods were implemented. The first was to use the relative time









consumed in queue (H) by each vehicle type. This method considered only the headway of each

vehicle i when following a passenger car plus the average additional headway of any vehicle j

trailing vehicle i. The results of this method are shown earlier in this chapter in Table 5-4.

The second was implemented adjusting the additional time consumed by each vehicle in

queue. This adjustment comes from the observation that the time consumed by a vehicle i it is

not exactly the same across different trailing vehicles. A general for of this method is shown in

Equation 5-19.




H d,_ = h,,PC + Ahl + [5-19]


Where;
Ha4J-l = adjusted time consumed by vehicle i in queue
hee = headway of vehicle i when following a PC
Ah, = average additional headway by vehicle i
h, = additional headway of vehicle i when following vehicle j

Using the same approach of relative time consumed, PCE values of 1.75, 2. 16, and 2.80

were obtained for small trucks, medium trucks, and large trucks respectively.

These values were extremely similar, and to be consistent with these results, it was decided

to use values of 1.8, 2.2, and 2.8 for small, medium and large trucks respectively. Tests done

with these values across a wide range of overall truck percentage, as well as relative truck type

percentages, showed that the resulting fH ValUeS tracked reasonably well with those yielded by

applying all 16 possible vehicle lead-follow PCE values. For situations where the relative truck

type percentages might be extremely skewed (such as an exit from a distribution warehouse that

consists entirely of large trucks) the specific PCE value from Table 5-2 can be used directly (in

this case, PCELT-LT).










Furthermore, when relative truck type distributions are not available, a general

approximation based on a relatively balanced distribution of small, medium, and large trucks in

the traffic stream can be assumed. This approximation consisted of the average of the PCEs for

these three categories. From Table 5-4 PCE values of 1.8, 2.2 and 2.8 can be obtained for small,

medium and large trucks respectively. The average of these values yields to 2.267. Thus, a

single truck PCE value of 2.3 can be applied. Again, this rough estimate considered a general

assumption that small trucks, medium trucks and large trucks are equally distributed in the traffic

stream.









CHAPTER 6
CONCLUSIONS AND RECOMMENDATIONS

The signal analysis methodology (Chapter 16) of the Highway Capacity Manual (HCM)

currently recommends a single truck PCE value of 2.0. There is not much research available to

support this value, and some transportation professionals have questioned the validity of this

value, feeling that it generally underestimates the impact of trucks on saturation flow. This issue

was the focus of this study; thus, the obj ective was to either validate this specific PCE value from

the HCM, or to recommend a more appropriate value, or combination of values.

The results of this study are based primarily on simulation data. While not as ideal as

having the results based strictly on field data, for this type of study that was focused on the

effects of large trucks, there were some additional constraints beyond the typical saturation flow

rate study that does not consider the effect of trucks. The challenge of finding sites with long

queues (eight or more vehicles) that also had a large percentage of trucks, the large variance of

the effect of trucks, and the sample size required due to this variance made it prohibitive to

collect sufficient field data in an efficient manner within the scope and resources of this project.

Nonetheless, a considerable amount of field data were still collected, which provided for a

reasonable data set to use for simulation program calibration. It was felt that the simulation

program was calibrated fairly well to the field data, and replicated the general trends observed in

the field data quite well. Thus, it is felt that the results provided in this study, despite the heavy

reliance on simulation data, are reasonably valid and reliable.

The results presented in this study were based on using the same definition for saturation

headway as provided in the HCM; that is, the average headway of vehicles in position 5 through

8 of the queue. Some studies [10] have indicated that the saturation flow rate will increase

(saturation headway will decrease) with longer queues; although the body of evidence at this










point is inconclusive. Li and Prevedouros [10] also found that in some cases the saturation

headway may increase, and in others it might decrease. They indicated that in the former,

headways at the end of a long queue compress, and in the latter headways elongate, and this is a

function of drivers' performance. In this study, it was found that trucks in the first few positions

of the queue impact not only the start-up lost time, but also the saturation headway of short

queues. This result likely reflects the fact that start-up lost time extends beyond the fourth

vehicle in queue, as it is expected that saturation headway would eventually reach a stable value

if the queue is sufficiently long enough. This result could not be confirmed directly, due to

computational limitations of the simulation program.

As indicated before, for purposes of this research it was assumed that the saturation section

of the queue started after the fourth vehicle. Therefore it was decided to exclude this increase in

the saturation headway because it is most likely a result of the additional lost time caused by the

trucks being absorbed into the saturation flow rate estimate for short queues. From a practical

standpoint, the accuracy of capacity estimates is not critical for situations when queues at

signalized intersections are consistently short.

As it was demonstrated, headways are a function of both the leading and following vehicle

in the queue, not just the trailing vehicle. This study categorized vehicles into four different

types, which led to a total of 16 (42) pOssible leading-trailing vehicle pairs. New PCE values

were developed for these pairs; however, it is not realistic to expect a practitioner to collect such

detailed data on vehicle pair frequencies to apply these PCE values. Therefore, in order to

provide more practical results, PCE values were developed based on three general categories of

truck type--small, medium, and large. The method by which this was accomplished was to

consider the time each vehicle type consumed during the queue clearance process. This time










consumed was defined as the headway of the vehicle plus an additional time it adds to the

trailing vehicle, and was based on the values obtained from the model that considered all 16

possible lead-trail vehicle pairs. The final recommended values for these PCEs are listed below.

* 1.8 for small trucks
* 2.2 for medium trucks
* 2.8 for large trucks

In this study small trucks include those trucks that have only two axles and between four

and six tires. It also includes passenger cars with a trailer and garbage trucks, regardless of their

number of axles. Medium trucks include those with three axles and usually range in length from

40 to 55 ft. Passenger cars with a trailer using the fifth-wheel, recreational vehicles (RVs) and

small trucks with a trailer are also included as medium trucks. Large trucks include those with

four or more axles, RVs with trailers and buses.

For situations in which only an estimate of the overall percentage of trucks in the traffic

stream is available, then the single truck PCE value of 2.3 can be applied. Again, this value is a

very general approximation, based a relatively balanced distribution of small, medium, and large

trucks in the traffic stream. Use of this single general truck PCE value is reasonable for planning

applications, where the level of precision may not be as important, or where it may not be

possible to know the percentages of different truck classifications in the traffic stream.

For situations where the truck percentages, by category, are very unbalanced, the values of

Table 5-2 can be used instead of the generalized ones. For example, if an approach to a

signalized intersection serves a warehouse distribution center, where almost all the trucks are

large, then the PCE value for LT-LT could be applied in the fw equation.

The increase in SLT due to trucks in the first four positions of the queue can be estimated

using the percentage of trucks in the traffic stream. This estimate was based on the assumption










that the trucks are distributed evenly throughout the queue. The results of this study showed a

SLT for passenger car-only queues of approximately 2.5 seconds, which is larger than the HCM

recommended value of 2.0 seconds. This value increases accordingly with an increase in truck

percentage, reaching a value of approximately 17.5 seconds for 100% large trucks in the traffic

stream.

Another finding of this study relates to the HCM recommended base saturation flow rate

value of 1900 pc/h/In. For the field data obtained in this study that included queues of only

passenger cars, this saturation flow was rarely ever reached. As the results indicated, the average

saturation headway for passenger car only queues was approximately 2.03 seconds, which

corresponds to a saturation flow rate of 1773 pc/hr. While it is certainly debatable as to whether

any of the field sites had what would be considered truly "ideal" conditions, per the general

HCM definitions, this was generally the case. Only under the most optimal driver behavior and

vehicle composition (e.g., only small passenger cars) conditions were average saturation

headways below 2.0 observed. Although this field data set was not extremely large, the HCM

recommended value for base saturation flow rate appears to be quite optimistic. This would

seem to be particularly the case where the percentages of SUVs, mini-vans, and pick-up trucks

currently make up a significant percentage of passenger "cars." Research by Kockelman and

Shabih [6] found average headway values for these vehicle types to be greater than those for

sedan-type passenger vehicles.

From detailed observations of the field data, it was found that a very high percentage of

discharging queues have one or more drivers that hesitate/lag during their start-up process.

While difficult to conclude exactly why this happens from the video recordings, it appears to

frequently be a result of general driver inattentiveness or distraction. That is, the phenomenon of









drivers "falling asleep at the wheel" while waiting to start-up from a stop at a signalized

intersection approach appears to be quite common--one or two drivers (within the first eight

positions of the queue) almost every cycle. These hesitating/lagging drivers have a significant

impact on the capacity of the intersection. Queues that contained obvious instances of this driver

hesitation were not included in the calculated value of 2.03; thus, this value still represents more

ideal driver behavior. Unfortunately, this "ideal" driver behavior, at least for an entire

discharging queue, seems to be quite rare. So based on just this phenomenon, the HCM

recommended base saturation flow rate of 1900 pc/hg/In would appear to be essentially

unattainable over any reasonable length of time.

Recommendation for further studies

Ideally, a very large field study should be conducted, including not only all the variables in

Bonneson' s study [13], but also a comprehensive consideration of trucks and other vehicle types

affecting the performance of the intersections. For example, heavily loaded trucks obviously

have poorer acceleration capabilities than an unloaded truck. However, it must be recognized

that trying to incorporate truck weight information into the methodology may not be practical,

due to the difficultly associated with obtaining this type of measurement. With respect to

determining the distribution of trucks within the queue (i.e., by queue position), a larger field

study might also provide for the ability to reach a statistically valid conclusion on this issue.

A study including longer queues should be done in order to better identify the length of

queue over which the Start-up Lost Time applies when trucks are present at the front of the

queue. From the results of this research, trucks in the first four queue positions impact the

discharge rate of vehicles in positions five through eight, and to better estimate the saturation

flow rate it was necessary to exclude this impact. Studying longer queues in the field will also

provide more insight into the issue of queue compression or elongation.










It is also recommended to study the frequency of inattentive drivers within the queue. It

was observed how these situations frequently resulted in large gaps between vehicles, therefore

resulting in an effective reduction of intersection capacity. This driver inattentiveness

phenomenon could be could be included as a factor in the adjusted saturation flow rate

calculation, somewhat analogous to a 'local adjustment factor'.











APPENDIX A
DATA COLLECTION EQUIPlVENT SETUP


Videor Camera


Signal Head


Mast Arm


Figure A-1. Data collection equipment setup for method 1








I ~


Traffic video and
signal status
recording VCR




-- Video monitor to

i check field setup











~Y~-~n ~- Signal Encoder





Figure A-2. Signal controller cabinet with data collection equipment installed for method 1











APPENDIX B
PICTURES OF VEHICLE TYPES BY CATEGORY


ecl


;illi
r


I*L. --

.r _


-"


ii
I;


_


Figure B-1. Small trcks. A) Panel truck. B) Garbage truck. C) Two-Axle Single-unit dump
truck. D) Small delivery truck. E) Passenger cars with trailers


r































.
b


~I~IF


;i II
I;
II


Figure B-1. Continued


.1































SC~ ~ _
~r*
""~*
~c, L*re


r~.i ...~._J
;ii ,


IYY~I


Y


- .. .~,..1..


Figure B-2. Medium trucks. A) Three-Axle Single-unit dump truck. B) Concrete Mixer. C)
Passenger car with trailer using fifth wheel. D) Delivery truck. E) Single-unit cargo
truck.


P


1
-~












"31~C ~~

Ir


~I _


~;
r ~- r*

j IL 1~.-


~ "?rr~-~


BEI


I ~I~-;-


r
;ir


~IF `~

*umr*y*,lr.- l;r.;~


B

Figure B-3. Large trucks. A) Tractor plus trailer.


B) Tractor plus flatbed. C) Buses.












I- -~ ---wwr*--- r.rHr*r*r


Figure B-3. Continued


II "


;









APPENDIX C
SIMULATION PROGRAM SCREEN SHOTS


Figure C-1. Simulation screenshot. A) Before signal turns green. B) Once the queue starts to
discharge











APPENDIX D

QUEUE COMPOSITION


Table D-1. Vehicle type per position in queue
Freq Position in queue
1st 2nd 3rd 4th 5th 6th 7th 8th
1 174 1 1 1 1 1 1 1 1
2 2 1 1 1 1 1 1 1 2
3 3 1 1 1 1 1 1 1 3
4 9 1 1 1 1 1 1 1 4
5 9 1 1 1 1 1 1 2 1
6 6 1 1 1 1 1 1 3 1
7 6 1 1 1 1 1 1 4 1
8 5 1 1 1 1 1 2 1 1
9 5 1 1 1 1 1 3 1 1
10 11 1 1 1 1 1 4 1 1
11 5 1 1 1 1 2 1 1 1
12 6 1 1 1 1 4 1 1 1
13 6 1 1 1 2 1 1 1 1
14 1 1 1 1 4 1 1 1 1
15 3 1 1 2 1 1 1 1 1
16 3 1 1 3 1 1 1 1 1
17 10 1 1 4 1 1 1 1 1
18 7 1 2 1 1 1 1 1 1
19 3 1 3 1 1 1 1 1 1
20 7 1 4 1 1 1 1 1 1
21 6 2 1 1 1 1 1 1 1
22 6 3 1 1 1 1 1 1 1
23 7 4 1 1 1 1 1 1 1
24 1 1 1 1 1 1 1 2 2
25 1 1 1 1 1 1 1 3 2
26 1 1 1 1 1 1 1 3 4
27 6 1 1 1 1 1 1 4 4
28 1 1 1 1 1 1 2 3 1
29 1 1 1 1 1 1 2 4 1
30 1 1 1 1 1 1 2 1 4
31 1 1 1 1 1 1 3 2 1
32 1 1 1 1 1 1 4 2 1
33 1 1 1 1 1 2 1 2 1
34 2 1 1 1 1 2 1 1 4
35 2 1 1 1 1 2 2 1 1
36 1 1 1 1 1 3 1 2 1
37 1 1 1 1 1 3 1 4 1
38 1 1 1 1 1 3 3 1 1
39 1 1 1 1 1 4 1 2 1
40 1 1 1 1 1 4 1 1 2
41 1 1 1 1 2 1 1 2 1
42 1 1 1 1 2 1 1 4 1
43 1 1 1 1 2 1 1 1 4
44 1 1 1 1 2 3 1 1 1
45 1 1 1 1 4 1 1 2 1
46 1 1 1 1 4 1 1 3 1
47 1 1 1 1 4 1 1 4 1
48 2 1 1 1 4 1 1 1 3
49 1 1 1 1 4 1 4 1 1
50 1 1 1 2 1 1 1 3 1
51 2 1 1 2 4 1 1 1 1
52 1 1 1 4 1 1 1 1 2
53 1 1 1 4 1 1 3 1 1












Table D-1. Continued
Case Freq Position in queue
1st 2nd 3rd 4th 5th 6th 7th 8th
54 1 1 1 4 1 4 1 1 1
55 3 1 1 4 4 1 1 1 1
56 1 1 2 1 1 1 1 3 1
57 2 1 2 2 1 1 1 1 1
58 1 1 2 4 1 3 1 1 1
59 1 1 3 1 3 1 1 1 1
60 1 1 4 1 1 1 1 1 4
61 1 1 4 1 1 1 3 1 1
62 2 2 1 1 1 1 1 4 1
63 1 2 1 1 1 1 1 1 4
64 1 2 1 1 1 2 1 1 1
65 1 2 1 4 1 1 1 1 1
66 1 2 2 1 1 1 1 1 1
67 1 2 3 1 1 1 1 1 1
68 2 2 4 1 1 1 1 1 1
69 1 3 1 1 1 1 2 1 1
70 1 3 1 1 2 1 1 1 1
71 1 3 1 1 3 1 1 1 1
72 1 3 2 1 1 1 1 1 1
73 2 3 3 1 1 1 1 1 1
74 1 4 1 1 1 1 1 1 4
75 1 4 1 1 1 4 1 1 1
76 1 4 1 3 1 1 1 1 1
77 1 4 2 1 1 1 1 1 1
78 2 1 1 1 1 1 4 4 4
79 1 1 1 1 1 3 3 1 2
80 1 1 1 1 2 4 1 1 4
81 1 1 1 1 3 1 3 2 1
82 1 1 1 1 4 1 1 4 2
83 1 1 1 1 4 4 1 1 4
84 1 1 1 2 1 2 1 2 1
85 1 1 1 3 1 4 1 1 4
86 1 1 1 3 1 4 3 1 1
87 1 1 1 4 4 4 1 1 1
88 1 1 2 1 3 1 1 1 3
89 1 1 2 1 4 1 2 1 1
90 1 1 4 1 1 1 1 4 2
91 1 1 4 2 1 1 1 3 1
92 1 1 4 4 1 4 1 1 1
93 1 2 4 1 1 1 4 1 1
94 1 2 2 1 1 1 2 1 1
95 1 2 2 1 3 1 1 1 1
96 1 2 2 3 1 1 1 1 1
97 1 3 1 1 1 1 2 2 1
98 1 3 1 1 1 4 4 1 1
99 1 3 1 1 3 2 1 1 1
100 1 3 4 1 1 1 1 1 4
101 1 3 4 1 1 1 2 1 1
102 1 4 1 1 1 1 1 4 4
103 1 4 2 1 4 1 1 1 1
104 1 1 1 1 1 4 4 4 4
105 1 1 1 4 4 1 4 1 4
106 1 1 3 4 1 4 1 1 2
107 1 1 3 4 4 4 1 1 1
108 1 2 1 1 3 3 4 1 1
109 1 4 1 4 1 1 1 3 4
110 1 4 2 1 4 1 1 4 1



















Table E-1. Headway statistics for 16 lead-trail vehicle combinations

Vehicle in position 2
Trailing 1 1 1 1 2 2 2 2 3 3 3
Leading l1 2 3 4 1 2 3 4 1 2 3
count 61 6 7 6 6 4 1 3 4 1 1
mean 3.171 4.217 5.060 6.516 3.836 6.413 5.050 8.157 3.644 4.440 7.920
stdev 0.844 0.744 0.801 0.967 0.972 2.423 1.608 0.464


344 4
412 3

0 6 2 2
4.457 6.238 6.130
0.627 3.115 0.170


~Cd
cM

clV

cl
clM
cl
ct


Vehicle in position 3
1 1 1 1 2 2
l 2 3 4 1 2

63 11 4 8 4 1
2.512 3.482 4.344 5.212 2.875 3.605
0.785 0.546 1.410 0.765 0.714


Vehicle in position 4
1 1 1 1 2 2
l 2 3 4 1 2

61 5 5 9 7 0
2.360 2.922 3.097 4.670 2.880
0.631 0.240 0.619 0.955 0.637


Vehicle in position 5
1 1 1 1 2 2
1 2 3 4 1 2

58 5 5 13 6 0
2.235 2.578 2.708 4.293 2.929
0.644 0.389 0.275 0.680 0.581


Trailing

Leading
count
mean
stdev




Trailing
Leading
count
mean
stdev




Trailing

Leading
count
mean
stdev


2
4

1
5.830






2
4

0







2
4


3 3
1 2

4 1
2.923 3.800
0.768


34 44 4
41 23 4

09 12 1
4.158 3.510 4.835 3.900
0.924 2.029


3
1

7
4.120
1.298




3
1


3 4 4
4 1 2

0 11 1
3.842 3.360
1.054




3 4 4
4 1 2

0 11 1
3.435 4.500
0.950


4
4

4
5.739
0.349




4
4

3
4.290
0.231


4.620


5.132 3.620 3.990
1.048

























































3.728

0.652


Table E-1. Continued

1 vehicle in position 6

Trailing 1 1 1 1 2 2

Leading 1 2 3 4 1 2
count 61 6 4 12 9 1

mean 2.297 3.548 2.833 3.973 3.147 3.165

stdev 0.509 0.673 0.383 0.747 1.135


3 3 4

3 4 1

2 1 6

5.195 4.240 5.331

1.520 1.653


2.876

0.975


4.000 5.345

0.714




4 4

3 4

0 2

5.895

1.407




4 4

3 4

2 4

4.925 5.377

3.231 0.900


1 vehicle in position 7
1 1 1

1 2 3

577 6

2.080 2.626 2.649

0.436 0.403 0.513


Trailing

Leading
Count

mean

stdev


2 2


4 4

1 2

10 1

3.807 4.490

1.575


2 3 4

12 1

3.070 3.040 5.090

0.566


6 8

4.288 2.691

0.647 0.686


8 1

3.239 7.590

0.826


1 1 2 2 2 2 3 3 3 3

3 4 1 2 3 4 1 2 3 4

6 7 5 1 1 2 3 0 0 0


2.701

0.715


1 vehicle in position 8
1 1

l 2

56 11

2.133 2.533

0.471 0.295


Trailing

Leading
count

mean

stdev


4 4

1 2

120O

3.933

1.116


3.010 6.970 3.350

0.269


3.982 2.676

0.549 0.438


















Table F-1. Simulation model calibration resul

Vehicle composition
Case 1 2 3 4 5 6 7
11 11 11 1 1
21 11 11 1 1
31 11 11 1 1
41 11 11 1 1
51 11 11 1 2
61 11 11 1 3
71 11 11 1 4
81 11 11 2 1
91 11 11 3 1
101 11 11 4 1


Field Simulation MSE calculation
8 2-8 5-8 2-8 5-8 2-8 5-8
1 2.36 2.18 2.39 2.05 0.001 0.018
2 2.46 2.64 2.50 2.25 0.002 0.155
3 2.66 2.59 2.60 2.41 0.003 0.031
4 2.80 2.69 2.75 2.67 0.002 0.000
1 2.61 2.56 2.57 2.38 0.001 0.034
1 2.97 2.68 2.74 2.66 0.050 0.000
1 2.92 3.19 3.01 3.14 0.009 0.003
1 2.47 2.44 2.58 2.40 0.012 0.002
1 2.75 2.94 2.76 2.70 0.000 0.061
1 3.13 3.33 3.03 3.18 0.009 0.022
1 2.57 2.44 2.60 2.41 0.000 0.001
1 2.81 3.04 3.06 3.22 0.059 0.034
1 2.42 2.29 2.60 2.24 0.033 0.003
1 2.94 2.86 3.09 2.82 0.023 0.002
1 2.77 2.40 2.62 2.09 0.022 0.092
1 2.52 2.09 2.81 2.13 0.085 0.002
1 3.03 2.39 3.12 2.20 0.007 0.034
1 2.70 2.38 2.62 2.06 0.007 0.100
1 2.71 2.24 2.85 2.15 0.020 0.007
1 2.89 2.11 3.15 2.21 0.071 0.010
1 2.50 2.18 2.64 2.10 0.019 0.007
1 2.70 2.28 2.87 2.15 0.030 0.016
1 2.81 2.18 3.21 2.20 0.159 0.000
2 2.69 2.70 2.69 2.57 0.000 0.016
2 3.21 3.33 2.85 2.87 0.129 0.206
4 2.89 3.42 3.03 3.18 0.021 0.057
4 3.03 3.42 3.24 3.54 0.042 0.013
1 3.35 3.97 2.91 2.96 0.196 1.024
1 2.97 3.57 3.15 3.39 0.032 0.033
4 3.31 3.51 2.91 2.97 0.156 0.288
1 2.75 2.66 2.95 3.04 0.040 0.142
1 3.36 4.30 3.22 3.51 0.019 0.612
1 3.00 2.96 2.78 2.73 0.050 0.052
4 2.93 3.19 2.91 2.96 0.001 0.049
1 2.89 3.02 2.77 2.72 0.015 0.092
1 3.05 3.05 2.96 3.04 0.008 0.000
1 3.26 3.83 3.29 3.63 0.001 0.038
1 2.97 3.50 3.07 3.24 0.010 0.070
1 2.93 3.10 3.25 3.57 0.100 0.215
2 2.86 3.08 3.17 3.42 0.094 0.120
1 3.30 2.77 2.79 2.57 0.255 0.042
1 3.12 2.91 3.17 3.24 0.002 0.111


1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 2
1 3
1 3
1 4
2 3
2 4
2 1
3 2
4 2
1 2
1 1
2 1
1 2
1 4
3 1
1 2
1 1
1 2
1 4


1 1 2
1 1 4
1 2 1
1 3 1
1 4 1


1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 1
1 2
1 2


APPENDIX F

CALIBRATION RESULTS AND MSE CALCULATIONS














Table F-1. Continued

Vehicle composition


Simulation


10SE calculation


4 5 6 7 8 2-8 5-8 2-8 5-8 2-8


2 1

2 3

4 1
4 1

4 1

4 1

4 1

1 1

4 1

1 1

1 1

1 4

4 1

1 1

1 1
1 3

3 1

1 1

1 1

1 1

1 1

1 2

1 1

1 1

1 1

1 1

1 1
2 1

3 1


1 4 2.72 2.50 2.93 2.81 0.044

1 1 2.84 2.54 2.93 2.82 0.010

2 1 2.69 2.81 3.26 3.12 0.324
3 1 3.85 3.70 3.40 3.36 0.201

4 1 3.41 2.96 3.55 3.60 0.020

1 3 3.43 3.44 3.25 3.11 0.032

1 1 3.32 3.66 3.54 3.61 0.051

3 1 2.49 2.44 2.93 2.64 0.190

1 1 2.86 2.66 3.24 2.73 0.141

1 2 2.97 2.65 3.24 2.42 0.075

1 1 3.67 2.63 3.42 2.74 0.059

1 1 3.59 3.08 3.58 3.00 0.000

1 1 3.65 2.83 3.57 2.62 0.006

3 1 2.93 3.23 2.95 2.63 0.000

1 1 2.81 2.28 2.81 2.08 0.000
1 1 3.53 3.45 3.57 2.70 0.002

1 1 2.89 2.29 3.15 2.39 0.067

1 4 3.09 2.25 3.37 2.59 0.080

1 1 2.96 2.66 3.47 2.74 0.260

4 1 3.40 3.79 3.21 3.06 0.036

1 4 2.91 2.74 2.95 2.63 0.002

1 1 2.40 2.48 2.83 2.41 0.182

1 1 2.96 2.32 3.27 2.16 0.093

1 1 2.61 1.89 2.83 2.08 0.047

1 1 3.08 2.09 3.00 2.10 0.006

1 1 3.05 2.37 3.28 2.16 0.054

1 1 2.74 2.88 3.06 2.48 0.103
1 1 2.86 1.92 3.06 2.27 0.039

1 1 2.65 1.92 3.17 2.39 0.280

1 1 2.82 2.18 3.07 2.13 0.064

1 1 3.42 2.35 3.17 2.09 0.062

1 4 3.50 3.51 3.44 2.60 0.004

1 1 3.08 3.19 3.68 3.02 0.363

1 1 2.93 1.69 3.51 2.13 0.335

1 1 3.16 2.13 3.39 2.16 0.053

4 4 4.36 5.62 3.73 4.41 0.403

1 2 3.76 4.06 3.19 3.46 0.316

1 4 3.40 4.01 3.43 3.71 0.001

2 1 2.97 2.72 3.28 3.31 0.097
4 2 3.09 3.00 3.65 3.81 0.319

1 4 3.21 3.55 3.78 4.02 0.327

2 1 3.28 3.03 2.98 2.72 0.091

1 4 2.81 2.80 3.57 3.45 0.574

1 1 3.00 2.97 3.63 3.58 0.399

1 1 3.43 2.94 4.03 3.42 0.364


0.097

0.084

0.096
0.114

0.410

0.108

0.003

0.039

0.004

0.053

0.012

0.005

0.044

0.363

0.042
0.576

0.010

0.116

0.008

0.525

0.012

0.006

0.027

0.037

0.000

0.044

0.162
0.123

0.223

0.002

0.069

0.824

0.032

0.195

0.001

1.462

0.363

0.092

0.347
0.647

0.216

0.092

0.434

0.372

0.227


4 1


11 11 1 4

11 11 3 3

11 12 4 1

11 13 1 3
11 14 1 1

11 14 4 1

11 21 2 1

11 31 4 1

11 31 4 3

11 44 4 1













Table F-1. Continued

Vehicle composition
Case 1 2 3 4 5

88 1 2 1 3 1

89 1 2 1 4 1

90 1 4 1 1 1
91 1 4 2 1 1

92 1 4 4 1 4

93 2 4 1 1 1

94 2 2 1 1 1

95 2 2 1 3 1

96 2 2 3 1 1
97 3 1 1 1 1

98 3 1 1 1 4

99 3 1 1 3 2

100 3 4 1 1 1

101 3 4 1 1 1

102 4 1 1 1 1
103 4 2 1 4 1

104 1 1 1 1 4

105 1 1 4 4 1

106 1 3 4 1 4

107 1 3 4 4 4

108 2 1 1 3 3
109 4 1 4 1 1

110 4 2 1 4 1


Field Simulation MSE calculation

2-8 5-8 2-8 5-8 2-8 5-8

2.84 2.44 3.13 2.70 0.088 0.067

3.09 2.49 3.44 3.06 0.120 0.330

3.56 2.92 3.72 3.19 0.027 0.075
3.30 2.57 3.68 2.75 0.141 0.033

3.06 2.46 4.06 2.90 0.995 0.190

4.37 3.39 3.73 2.96 0.415 0.182

3.38 2.10 3.00 2.36 0.145 0.066

3.83 2.46 3.16 2.37 0.451 0.008

2.78 2.32 3.15 2.08 0.140 0.059
2.88 2.57 3.25 2.81 0.139 0.057

3.65 3.61 3.82 3.81 0.030 0.043

3.38 2.97 3.35 2.71 0.001 0.071

3.21 2.34 3.65 2.55 0.195 0.044

3.19 2.43 3.60 2.48 0.171 0.003

3.59 3.42 3.91 3.45 0.106 0.001
3.95 2.83 3.86 2.59 0.007 0.055

4.07 4.82 4.21 5.24 0.019 0.172

3.66 3.28 4.27 3.83 0.375 0.307

4.03 3.04 3.96 3.15 0.006 0.012

3.58 3.04 4.27 3.29 0.471 0.059

3.76 3.80 3.78 3.80 0.000 0.000
3.30 2.12 4.23 3.10 0.868 0.972

4.58 3.87 4.30 3.37 0.074 0.247









LIST OF REFERENCES


1. Highway Capacity Manual. Transportation Research Board (TRB), National
Research Council, Washington, D.C., 2000.

2. Washburn, S.S., Courage, K.G., and MacKenzie, S. Development of a Red Light
Violation Data Collection Tool. Southeastern Transportation Center, Final Report.
2001. http://stcuutk edu/htm/pdf%/20fil es/RLRTRB1. pdf

3. Highway Capacity Manual. Highway Research Board, Special Report 87, National
Research Council, Publication 1328, Washington D.C., 1965.

4. Molina, CJ, Jr. Development of Passenger Car Equivalencies for Large Trucks at
Signalized Intersections. Institute of Transportation Engineers, Journal vol. 57 Issue
11, pp. 33-37. November 1987.

5. Benekohal, R and Zhao, W. Delay-Based truck Equivalencies at Signalized
Intersections: Results and Field Data. Third Intemnational Symposium on Highway
Capacity, Copenhagen, Denmark. June 1998.

6. Kockelman, K and Shabih, R. Effect of Light-Duty Trucks on the Capacity of
Signalized Intersections. Journal of Transportation Engineering: American Society of
Civil Engineers, Volume: 126 Issue: 6 pp. 506-512. November 2000.

7. Bonneson, JA and Messer, CJ. Phase Capacity Characteristics from Signalized
Interchange and Intersection Approaches. Transportation Research Record: Joumnal
Transportation Research Board. No. 1646. TRB, National Research Council.
Washington D.C., pp. 96-105. 1998.

8. Bonneson, JA, Nevers, B., Zegeer, J., Nguyen, T. and Fong, T. Guidelines for
Quantifying the Influence of Area Type and Other Factors on Saturation Flow Rate.
Final Report #DO23 19. Florida Department of Transportation. Tallahassee, FL. June 2005.

9. Perez-Cartagena, R. and Tarko, A. Calibration of Capacity Parameters for Signalized
Intersections in Indiana. 84th Annual Meeting of the Transportation Research Board,
January 9-13, 2005, Washington D.C.

10. Li, H. and Prevedouros, P. Detailed Observations of Saturation Headway and Start-
Up Lost Times. Journal of the Transportation Research Board, TRR 1802,
Washington, D.C., 2002.

11. Cohen, S. L. Application of Car-Following Systems to Queue Discharge Problem at
Signalized Intersections. Transportation Research Record: Joumnal of the
Transportation Research Board, TRR 1802, Washington, D.C., 2002.

12. Greenshields, B. D. Distance and Time Required to Overtake and Pass Vehicles.
Highway Research Board Proceedings, Vol. 15, 1935, pp.332-342.










13. Bonneson, James A. Modeling Queued Driver Behavior at Signalized Junctions.
Transportation Research Record: Journal Transportation Research Board. No. 1365.
TRB, National Research Council. Washington D.C., pp. 99-107. 1992.

14. Briggs T. Time Headways on Crossing the Stop Line after Queueing at Traffic
Lights. Traffic Engineering, July & Control, May 1977, pp. 264-265.

15. Messer C. J. and Fambro D. B. Effects of Signal Phasing and Length of Left-Turn
Bay on Capacity. In Transportation Research Record 644, TRB National Research
Council, Washington, D.C., 1977, pp. 95-101.

16. Buhr J.H., Whiston R.H., Brewer K.A., and Drew D.R. Traffic Characteristics for
Implementation and Calibration of Freeway Merging Control Systems. In Highway
Research Record 279, HRB, National Research Council, Washington, D.C., 1969, pp.
87-106.

17. Evans L. and Rothery R.W. Influence of Vehicle Size and Performance on
Intersection Saturation Flow. Proc., 8th International Symposium on Transportation
and Traffic Theory, University of Toronto Press, Toronto, Ontario, Canada, 1981, pp.
193-222.

18. George E.T. and Heroy F.M. Starting Response of Traffic at Signalized Intersections.
Traffic Engineering, July 1966, pp. 39-43.

19. Akgelik R., Besley M. and Roper R. Fundamentals Relationships for Traffic Flows at
Signalised Intersections. Research Report ARR 340. ARRB Transport Research Ltd,
Vermont South, Australia. 1999

20. Pipes L.A. An Operational Analysis of Traffic Dynamics, Journal of Applied Physics,
Vol. 24, NO. 3, 1953, pp. 274-287.

21. May, A. Traffic Flow Fundamentals. Published by Prentice-Hall, Inc., 1990, Chapter
6, pp. 160-191.

22. Long, Gary. Intervehicle Spacing and Queue Characteristics. Transportation
Research Record: Journal Transportation Research Board. No. 1796. TRB, National
Research Council. Washington D.C., pp. 86-96. 2002.

23. Florida Department of Transportation. ARTPLAN Level Of Service analysis
software. http://www.dot. state.fl.us/planning/systems/sm/los/los_sw.t

24. Cohen, S. L. Application of Car-Following Systems in Microscopic Time-Scan
Simulation Models. Journal of the Transportation Research Board, TRR 1802,
Washington, D.C., 2002.

25 StatSoft, Inc. (2006). STATISTICA (data analysis software system), version 7.1
www. statsoft. com.









BIOGRAPHICAL SKETCH

Carlos Cruz-Casas is a 24-year-old graduate student at the University of Florida. He

completed a Master of Engineering degree in transportation engineering. He received a Bachelor

of Science in civil engineering degree from the University of Puerto Rico-Mayagiiez in May of

2005.





PAGE 1

1 DEVELOPMENT OF PASSENGER CAR EQUIVALENCY VALUES FOR TRUCKS AT SIGNALIZED INTERSECTIONS By CARLOS O. CRUZ-CASAS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ENGINEERING UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Carlos O. Cruz-Casas

PAGE 3

3 To my parents, for their endless support

PAGE 4

4 ACKNOWLEDGMENTS I would like to acknowledge the excellent the guidance and support of the committee chair, Dr. Scott S. Washburn, and committee members Dr Lily Elefteriadou and Dr. Yafeng Yin. I would also like to acknowledge Diego Arguea an d Tom Hiles, for their significant collaboration in this research. In addition, I would like to thank Matthew O Brien, Darrell Schneider, and the City of Gainesville for their support du ring the field data collection process.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES................................................................................................................ .........8 ABSTRACT....................................................................................................................... ............10 CHAPTER 1 INTRODUCTION................................................................................................................. .12 Background..................................................................................................................... ........12 Problem Statement.............................................................................................................. ....13 Research Objective and Tasks................................................................................................14 Document Organization.......................................................................................................... 14 2 LITERATURE REVIEW.......................................................................................................16 Overview of HCM Treatme nt of Heavy Vehicles..................................................................16 Passenger Car Equivalency Factor at Signalized Intersections..............................................18 Car-Following Models........................................................................................................... .23 Summary of HCM 2000 PCE Guidelines...............................................................................31 Summary of Previous PCE Studies........................................................................................31 Summary of Car-Following Models.......................................................................................32 3 RESEARCH APPROACH.....................................................................................................34 Methodological Approach......................................................................................................34 Field Data Collection.......................................................................................................... ....37 Simulation Modeling............................................................................................................ ..49 4 ANALYSIS OF FIELD DATA AND CA LIBRATION OF SIMULATION MODEL.........56 Summary of Data Reduction..................................................................................................56 Time Headways.................................................................................................................. ....57 Start-up Reaction Time (SRT)................................................................................................60 Start-up Lost Time (SLT)....................................................................................................... 62 Calibration of Simulation Model............................................................................................63 Execution of Experimental Design.........................................................................................65 5 ANALYSIS OF SIMULATION DATA................................................................................67 6 CONCLUSIONS AND RECOMMENDATIONS.................................................................84

PAGE 6

6 APPENDIX A DATA COLLECTION E QUIPMENT SETUP......................................................................90 B PICTURES OF VEHIC LE TYPES BY CATEGORY...........................................................92 C SIMULATION PROGRAM SCREENSHOTS......................................................................97 D QUEUE COMPOSITION......................................................................................................98 E HEADWAY STATISTICS..................................................................................................100 F CALIBRATION RESULTS AND MSE CALCULATIONS..............................................102 LIST OF REFERENCES............................................................................................................. 105 BIOGRAPHICAL SKETCH.......................................................................................................107

PAGE 7

7 LIST OF TABLES Table page 2-1 Car-following models comparison.....................................................................................32 3-1 Possible pairing combinati ons for four vehicle types*......................................................36 3-2 Data collection cites...................................................................................................... .....38 3-3 Data collection periods for Willis ton Rd/34th Street site (method 1)...............................48 3-4 Data collection periods for other sites (method 2).............................................................48 4-1 Average headway and frequencies for each leader-follower combination........................57 4-2 Time headways from field data..........................................................................................58 4-3 Results from STATISTICA for model 1...........................................................................58 4-4 Results from STATISTICA for model 2...........................................................................59 4-5 Results from STATISTICA for model 3...........................................................................59 4-6 Results from STATISTICA for model 4...........................................................................60 4-7 Distribution of trucks by site and position in queue..........................................................62 4-8 Final choice of calib ration parameter values.....................................................................65 5-1 Model estimation results for additional h eadways of 16 vehicle pair combinations.........73 5-2 Headways and PCE values fo r 16 vehicle pair combinations............................................75 5-3 Vehicle types headway and their h ................................................................................76 5-4 Time consumed and PCE values for each vehicle type.....................................................77 5-5 Results from STATISTICA fo r model with vehicle types................................................78 5-6 PCE factors for three vehicle types....................................................................................79 5-7 Model estimation results for fHV equation with 16 vehicle pairs.......................................81 D-1 Vehicle type per position in queue.....................................................................................98 E-1 Headway statistics for 16 l ead-trail vehicle combinations..............................................100 F-1 Simulation model calibration results...............................................................................102

PAGE 8

8 LIST OF FIGURES Figure page 2-1 Measured headways including start-up lost time (dark shade)..........................................18 3-1 SW Williston Rd / SW 34th St ae rial view (eastbound on Williston Rd).........................38 3-2 SW Williston Rd / SW 34th St ground level view (eastbound on Williston Rd)..............39 3-3 University Ave. / Waldo Rd aerial view (northbound on Waldo Rd)...............................39 3-4 University Ave. / Waldo Rd gr ound level view (northbound on Waldo Rd)....................40 3-5 US 41 / SR 50 aerial view (westbound on SR50)..............................................................40 3-6 US 41 / SR 50 ground level view (westbound on SR50)...................................................41 3-7 US 301 / SR 50 aerial view (eastbound on SR 50)............................................................41 3-8 US 301 / SR 50 ground level view (eastbound on SR 50).................................................42 3-9 CR 326 / SR 200A ground level view (westbound on CR 326)........................................42 3-10 John Young Parkway / Colonial Dr ae rial view (southbound on John Young Pkwy)......43 3-11 John Young Pkwy / Colonial Dr gr ound level view (southbound on John Young Pkwy).......................................................................................................................... .......43 3-12 Preferred video camera mounting location (plan view) for method 1...............................45 3-13 Screen capture (low resolution) of video image from Williston Rd/34th Street site in Gainesville (method 1).......................................................................................................45 3-14 Camera setup for data collection (method 2).....................................................................46 3-15 Screen capture of the composite video image (method 2).................................................47 3-16 Simulation program user interface.....................................................................................52 4-1 Start-up reaction time frequencies.....................................................................................61 4-2 Start-up reaction time freque ncies for trimmed data set....................................................61 4-3 User-adjustable vehicle, driver, and model parameters.....................................................64 5-1 Queue sections for analysis................................................................................................ 67 5-2 Impact of trucks in the first part of the queue to the saturation headway..........................68

PAGE 9

9 5-3 SLT example................................................................................................................ ......72 5-4 Total time for vehicles 1-8 using vehicle pairs in positions 5-8........................................74 5-5 Time consumed by a large truck........................................................................................75 5-6 Total time for vehicles 1-8 us ing vehicle types in positions 5-8.......................................78 A-1 Data collection equipment setup for method 1..................................................................90 A-2 Signal controller cabinet with data co llection equipment installed for method 1..............91 B-1 Small trucks. A) Panel truck. B) Ga rbage truck. C) Two-Axle Single-unit dump truck. D) Small delivery truck. E) Passenger cars with trailers.......................................92 B-2 Medium trucks. A) Three-Axle Single-u nit dump truck. B) Concrete Mixer. C) Passenger car with trailer using fifth wheel D) Delivery truck. E) Single-unit cargo truck.......................................................................................................................... .........94 B-3 Large trucks. A) Tractor plus trailer. B) Tractor plus flatbed. C) Buses........................95 C-1 Simulation screenshot. A) Before signal turns green. B) Once the queue starts to discharge...................................................................................................................... ......97

PAGE 10

10 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Engineering DEVELOPMENT OF PASSENGER CAR EQUIVALENCY VALUES FOR TRUCKS AT SIGNALIZED INTERSECTIONS By Carlos O. Cruz-Casas August 2007 Chair: Scott S. Washburn Major: Civil Engineering Large trucks have considerably different size and performance ch aracteristics than passenger cars. Consequently, these trucks can have a significant impact on traffic operations. It is therefore essential to properly account for this impact in the tra ffic operations analysis in order to reflect the operational quality of the road way as accurately as possible. Signalized intersections are one roadway facility that can be particularly sensitive to the presence of commercial truck traffic. The most common method used for the analysis of signalized intersections is contained in the Highway Capacity Manual (HCM). In this method, the base saturati on flow rate of the signalized intersection is defined in units of passenger cars per hour green per lane (pc/hg/ln). To account for the presence of large trucks in the traffic stream, the HCM includes a Passenger Car Equivalency (PCE) value. In the current ed ition of the HCM, a PCE value of 2.0 is applied for all large trucks, with no distinctio n between different sizes of trucks. Some transportation professionals have que stioned the validity of this PCE value recommended by the HCM. They are concerned that the impact of tr ucks at signalized intersections is being unde r-estimated. If this is the case, then capacity is being over-estimated and intersections are not be ing adequately designed.

PAGE 11

11 The primary objective of this research was to determine appropriate truck PCE values to apply for signalized intersection analysis. Thes e PCE values were classi fied by three different categories of truck sizes and pe rformances. Additionally, a gene ral PCE value with only one truck category was developed for planning purpo ses and/or a less detailed analysis. The development of the PCE values was based on the relative headway concept, as defined in the HCM. The results of this study are based primar ily on data generated from a custom simulation program. However, a considerable amount of field data was collected for the purpose of simulation calibration. The PCE values determin ed from this study are 1.8, 2.2, and 2.8 for small trucks, medium trucks, and large trucks, respec tively. Additionally, an equation was developed to calculate start-up lost time that accounts for the impact of trucks at the front of the queue, as opposed to the standard 2.0 seconds recommended by the HCM. Furthermore, based on the field data collected, it was found that the base sa turation flow rate va lue of 1900 pc/hg/ln recommended by the HCM appears to be quite optimistic.

PAGE 12

12 CHAPTER 1 INTRODUCTION Background Large trucks have considerably different size and performance ch aracteristics than passenger cars. Consequently, these trucks can have a significant impact on traffic operations. It is therefore essential to properly account for this impact in the tra ffic operations analysis in order to reflect the operational quality of the roadway as accurately as possible. This study focuses only on those trucks that are consider ably larger than pick-up trucks. Signalized intersections are one roadway facility that can be particularly sensitive to the presence of commercial truck tra ffic. Like other fac ilities, the length of trucks has a negative impact on the capacity of the signalized in tersection; however, th e reduced performance characteristics of these trucks has an even greater impact on signalized intersection than uninterrupted flow facilities du e to the need for many of th e trucks traveling through an intersection to decelera te to a stop and re-accel erate to cruise speed. Given the longer time it takes trucks to re-accelerate to cruise speed, when compared to passenger cars, the presence of trucks can have implications on si gnal coordination. This effect may be greater when trucks are present at the front of a discharging queue. When the traffic stream stops at the inte rsection, the inter-vehicle spacing decreases, resulting in increased vehicle dens ity. At this higher density it is clear that trucks occupy more space than passenger cars due to their physical char acteristics. Once the signal turns green, and after a short period of start-up lost time, vehicles start departing at the saturation flow rate and it is here where the heavy vehicles have the great er impact due to their operational capabilities. Large trucks have poorer acceleration than passenger cars; therefore it will take them more time to reach their desired speed. With poorer acceler ation characteristics, he avy vehicles slow down

PAGE 13

13 the traffic stream and increase their time headway. Furthermore, with their longer length, heavy vehicles increase the time headway of the vehicle following them. In arterial roads where the Free Flow Speed is low, the cruising speed of a ll vehicles may be similar. Once the vehicles reach their desired speed and have a constant crui sing speed, the impact of trucks to the traffic will be again mostly due to their length. Problem Statement The most common method used for the analysis of signalized intersections is contained in the Highway Capacity Manual (HCM) [1]. In this method, the base saturation flow rate of the signalized intersection is defined in units of passenger cars per hour green per lane (pc/hg/ln). To account for the presence of large trucks in the traffic stream, the HCM includes a Passenger Car Equivalency (PCE) value. This factor is base d on the relative headway of a truck to that of a passenger car. In the current edit ion of the HCM, a single PCE valu e of 2.0 is used for all trucks passing through a signalized intersection. Thus, this implies that a single large truck is equivalent to two passenger cars for capacity anal ysis purposes; that is, an intersection can only accommodate half as many trucks as cars. Transportation professionals w ith the Florida Department of Transportation (FDOT) have recently questioned the validity of this PCE valu e recommended by the HCM. This concern has stemmed from observations that the failure rate of intersections in Florida seems particularly high when any significant percentage of trucks is present in the traffic stream. Their observations are especially distur bing in light of the fact that Floridas populat ion is growing rapidly, and complimentary to this population gr owth is a corresponding growth in commercial truck traffic. Thus, FDOT offi cials are concerned that the imp acts of trucks at signalized intersections in Florida are being under-estimated. If this is the case, then capacity is being overestimated and intersections are not being adequately designed.

PAGE 14

14 Research Objective and Tasks The primary objective of this research was to determine appropriate truck PCE values to apply for signalized intersection analysis. Thes e PCE values were classi fied by three different categories of truck sizes and pe rformances. Additionally, a gene ral PCE value with only one truck category was developed for planning purpose s and/or a less detailed analysis. The results of this study are based primarily on data ge nerated from a custom simulation program. However, a considerable amount of field data was collected for the purpose of simulation calibration. The following tasks were conducted to suppor t the accomplishment of these objectives: Conduct a literature review Collect preliminary field data Develop data collection methods Determine appropriate criteria to select data collection sites Identify appropriate field data collection sites Collect video data from field Process the video using the RLRAP [2 ] to obtain signal status and time Reduce field data Develop a simulation program Calibrate the simulation program Develop an experimental design Generate simulation data se t from experimental design Perform analysis and modeling of the generated simulation data set Develop new PCE factors Document Organization Chapter 2 presents an overview of the relevant studies found in the liter ature. This review includes the state of the practice with regard to incorporating the effects of large vehicles into traffic analyses, previous studies focused on the development of PCE values on arterial roads or at intersections, and a review of car-following models that might be applicable to modeling queue discharge at signalized inte rsections. Chapter 3 describes th e research approach that was used to accomplish the objectives of this st udy including the methodological approach, field data

PAGE 15

15 collection, simulation model development, and th e simulation experiments. Chapters 4 and 5 contain the analysis and results of the field and simulation data respectively. Conclusions and recommendations are contained in Chapter 6.

PAGE 16

16 CHAPTER 2 LITERATURE REVIEW An extensive literature review has been conducted in three areas. The first area is a review of the state of the practi ce with regard to incorporating the e ffects of heavy vehicles into traffic analyses, namely the Highway Capacity Manual [1, 3]. The second area targets previous studies that focused on the development of PCE values at intersections. While ot her studies have been done on the development of PCEs for other facilities, they were not covered in this project due to the unique influence of signals on traffic flow. A custom simu lation program was developed for use in this project for several r easons, which are discussed in Chap ter 3. Therefore the third area of this chapter deals with a revi ew of car-following models that might be applicable to modeling queue discharge at signalized intersections. Overview of HCM Treatment of Heavy Vehicles The Highway Capacity Manual (HCM) firs t introduced the term passenger car equivalent in the 1965 version of this publication [3] as the numb er of passenger cars displaced in the traffic flow by a truck or bus, under the prevailing roadway and traffic conditions. The HCM 2000 [1] definition states that the passenger car equivalent is the number of passenger cars displaced by a single, heavy vehicle of a part icular type under specified roadway, traffic, and control conditions. Currentl y, a PCE value of 2.0 is specified for all heavy vehicles. The refining of the original definiti on emphasizes the importance of tra ffic controls and their effect on PCE values, and considers the subjectivity of wh at can be considered as heavy vehicles by not specifically citing those vehicles. As defined by the HCM 2000, a heavy vehicle is any vehicle which has more than four tires in contact with the drivi ng surface. There is no distinc tion between trucks, recreational vehicles, and buses in the calcu lation of the adjusted satura tion flow rate at signalized

PAGE 17

17 intersections. The HCM al so recommends that if no data ex ist for a particular intersection, a value of 2% heavy vehicles should be used for urban streets. The PCE is implemented through a heavy-vehicle factor ( fHV), which is used to adjust the base saturation flow rate. This heavyvehicle factor is one of seve ral adjustment factors for the base saturation flow rate ( S0). In Equation 2-1 is shown how the base saturation flow rate is adjusted by the as it is defined in Equation 16-4 of the HCM 2000. HVf S S 0 [2-1] The form of the equation for fHV (Equation 2-2) is: 1 1 1 T T HVE P f [2-2] Where: PT = percentage of trucks in the traffic stream ET = passenger car equivalency factor The standard procedure for measuring satu ration flow rate (HCM2000, 16-158 in appendix H) prescribes that the headways of the first four to six vehicles in queue are not considered as saturation headway because this time is usually considered to be pa rt of the start-up lost time. The HCM 2000 recommends a default start-up lo st time (SLT) of 2.0 seconds if field measurements are not available. It is not specif ied for what traffic stream composition this value is based upon (e.g., passenger car only stream). This start up lost time can be determined by adding the difference between saturation headway and the headway measured for those first few vehicles that do not depart under the saturation headway. Figure 2-1 illustrates this con cept, where the saturation headwa y has a value of 2.0 seconds per vehicle (light shade) and the lost time is shown in this example as applying to the first four vehicles (dark shade).

PAGE 18

18 0 0.5 1 1.5 2 2.5 3 3.5 12345678 Vehicle in QueueHeadway, (s) Figure 2-1. Measured headways incl uding start-up lost time (dark shade) Passenger Car Equivalency Factor at Signalized Intersections In 1987, Molina [4] derived a PCE model ba sed on the assumption that passenger cars depart at a constant saturation flow headway, and thus the basis for his model is the headway method. Molina collected field data from one site in each of three cities in Texas, and obtained a total of 13,000 observations. During the data coll ection, Molina considered vehicles to form a part of the queue if they came to a complete, or near stop. He recorded the time that these vehicles crossed the stop line. He classified the 13,000 vehicle observatio ns into four vehicle classes and within each vehicle class into ten queue positions He used the regression analysis method to develop a model of the collected data An expression was derived for the additional effect of a heavy vehicle in the first position of a queue. This is a modified expression of the headway ratio method, and derive d relationships consider only one heavy vehicle in the queue with its position varying from one to ten. In addition, the analysis was limited to through movements only, and other factors, such as percentage of trucks, vehicular volumes, and head way increase of the ei ghth-positioned vehicle

PAGE 19

19 behind the truck are not consid ered. Molina found that posi tion in queue did not have a pronounced effect when dealing with twoand th ree-axle and single-unit trucks, but had a very pronounced effect with five-axle combination trucks. His recommendations included using different methods to distinguish between light and heavy vehicles when analyzing capacity at signalized intersections, as th ese truck types can have sign ificantly different effects. Benekohal and Zhao [5] performed a study on the additional delay to the passenger cars behind a heavy vehicle. This delay is produced from both longer headways as well as additional headway increases of those vehicles behind the he avy vehicle that causes the delay. This study introduces a new PCE value labeled the D-PCE, meaning a delay-based calculation of passenger car equivalents. The delay-based passenger car equivalent is defined by Benekohal and Zhao as the ratio of delay caused by a heavy vehicle to the delay of a car in an all-passenger car traffic stream. The calculation of these D-PCE values considers the traffic volume as well as the percentage of heavy vehicles in the traffic str eam. Data were collected at ten approaches of seven intersections in Central Illinois, where sites possessed as many ideal features as possible. Vehicle headways, delays for all-passenger ca r streams, number of queued and non-queued vehicles, the position of heavy vehicles in the queue, signal timing information, and geometric data were all collected in the process. The h eadway time of the first vehicle was defined from the moment the signal turned green to the point when the rear wheels of the vehicle crossed the stop line. This implies that reaction tim e was included in the start-up lost time. Because TRAF-NETSIM queue delay was used for comparison, queue delay was the performance measure recorded in the field. Be nekohal and Zhao concluded that the position of the truck in queue is not as important as how many vehicles are behind the truck. Also, comparison with the HCM values for single unit tr ucks indicates that the HCM overestimates the

PAGE 20

20 effect by which capacity is reduced by these types of vehicles at signalized intersections. The DPCE increases with the number of vehicles behind a heavy vehicle, and PCE values for signalized intersections should be determined ba sed on additional delay caused by large trucks. Kockelman and Shabih [6] performed a study of the impact of light-duty trucks (LDTs) on the capacity of signalized intersections. Three factors were identified as influencing vehicle headways: length, performance, and driver beha vior. Kockelman and Shabih used the headway method to arrive at PCE values for five diffe rent categories of light -duty trucks, where the additional time it takes for a passenger car behind an LDT to enter the intersection relative to being behind a passenger car is considered. The start-up lost time in th eir developed model did not include the reaction time of the first driver as measurements began only when the first vehicle began to move (i.e., they did not record the start of green). In addition, only those vehicles that came to a complete stop before the signal changed to green we re considered to have been part of the queue. Field data were collected from sites that met the following criteria: High traffic volumes a nd significant queuing Level terrain Exclusive left turn lane and protected signal phase for left turns Exclusive right turn lane Ease of data collec tion equipment setup Mix of vehicle types No parking zones along streets Insignificant disturbance from bus stops The time elapsed from when the first vehicle in the queue began to move to when the rear axle of the last vehicle in the queue crossed the stop line was measured. After a statistical analysis, Kockelman and Shabih concluded that vehicle length is a significant factor on following-vehicle headways. As a result, th e impacts of LDTs should be given special consideration, as sport-utility vehicles as well as vans have the ability to reduce the capacity at a signalized intersection in a statistically significant manner. Kockelman and Shabih

PAGE 21

21 recommended PCE values of 1.07, 1.41, 1.34, and 1.14 for small SUV, long SUV, vans and pick-up trucks respectively. Bonneson and Messer [7] performed a study in wh ich several models were developed that can be used to predict the satu ration flow rate and start-up lo st time of through movements at signalized interchange ramp terminals and othe r closely spaced intersections. The minimum discharge headway method is used, and it is typi cally reached by the vehicle in the sixth position of the queue. In their models they include the te rm traffic pressure, defined as the tendency for vehicle headways to decrease as queue lengths increase and aggressive driving is present. Bonneson and Messer indicate that other authors (Stokes et al.) have independently identified this occurrence and call it headw ay compression. Data for the study were collected at twelve interchanges in five states. Th e sites selected for analysis co ntained the two basic forms of interchanges, partial cloverle af and diamond. Video cameras and computer-monitored tape switch sensors, mounted at the upstream end of each of two street segments, were used for data collection. The data were collected during weekdays, betw een 7:00 a.m. and 7:00 p.m. The study concluded that there exists a strong correlation between start-up lost time and saturation flow rate, and that the distance to a do wnstream queue as well as traffic pressure has a significant effect on the saturation flow rate of a signalized traffi c movement. Therefore, start-up lost time is not a constant value as it is comm only used in practice, but rather dependent upon the saturation flow rate. Bonneson and Messer recommended that an id eal saturation flow rate of 2000 pc/h/ln should be used for high vol ume intersections in urban areas. Bonneson et al. [8] studied the formulati on of the equation 16-4 of the HCM 2000 for saturation flow rate at intersections. The research resulted in the development of some new adjustment factors such as ar ea population, number of lanes, and right-turn radius and the

PAGE 22

22 revision of some existing factors such as right tu rn, traffic pressure, and heavy vehicle. These adjustment factors were develope d from data collected five differe nt counties in Florida. Data were collected at 12 intersections, which include d a total of 38 approaches and 2901 cycles in overall. While not a focus of this study, a truck PCE factor was included in the model, as some trucks were present in the data set. Their es timated PCE value was 1.74, which interestingly is smaller than the HCM recommended value of 2.0. However, the authors indicate that truck percentages were not significant in their data set, and they specifically recommend a more thorough investigation of this specific factor. Perez-Cartagena and Tarko [9] in their study focu sed in the development of local values of the base saturation flow rate and lo st times used in capacity analysis of signalized intersections in Indiana. In their study, it was considered that th e first four vehicles in queue were carrying the SLT. They found that some of the default va lues recommended by the HCM were adequate for their location. These factors in cluded the heavy vehicle factor ( fHV). They also found that the saturation flow rate was not the sa me for all their sites even though they were almost identical in terms of geometry and traffic conditions. This indicated that there are some other factors affecting the saturation flow rate that are not considered in the HCM. Perez-Cartagena and Tarko proposed population adjustment factors of 0.92 for medium towns and 0.79 for small towns to be added to the adjusted sa turation flow equation 16-4 in HCM 2000. Li and Prevedouros [10] exam ined saturation headway and st art-up lost times of traffic discharging from a signalized inte rsection. Their study was done using data collected from one through movement and one protected left turn at a single intersection. They proved that the assumption of that the saturation headway of s hort and long queues is th e same was overlooking

PAGE 23

23 other factors that might be present. It was obser ved how the last few vehicles in a longer queue can produce either compressed or elongated headways. Compressed headways were observed when vehicl es bunch together to be able to cross the stop bar before the clearance in terval is over. Elongated headways were observed when the queues were long enough to allow the vehicles to exceed speeds of 40 mph. Additionally, they found that the minimum headway was not reached until the 9th to 12th vehicle in queue. In addition Li and Prevedouros recommended a mean start-up reaction time of 1.76 seconds with a standard deviation of 0.61. Car-Following Models Cohen investigated the issue of simulating queue discharge at a signalized intersection through the application of the modified Pitt car-following model [11]. This study was based on a single intersection with no rest rictions on the departing fl ow. Some queue-discharge mechanisms are based on the assumption of every vehicle in queue departing from the intersection at equal time headways. These assumptions are neglecting the start-up delay brought from the first few cars in queue and othe r issues such as varying vehicle and driver characteristics, among others. The basic form of the modified Pitt car-f ollowing model is shown in Equation 2-3. T h T T R t a T R t v R t v v h L R t s R t s K T t al l f f l f l f2 1 2 12 [2-3] This model estimates an acceleration for a follo wing vehicle, subject to three constraints: af has to be between amin and amax; the speed at any time t has to be less than the free-flow speed; and af has to be less than the acceleration co mputed for safe following (Equation 2-4):

PAGE 24

24 min 2 min 22 5 0l l l f l f fea T t v L T t s T t s R t v T t a [2-4] In these two equations the variables are defined as: af( x ) = acceleration trailing vehicle at time x computed from car-following (ft/s2) t = current simulation time (sec) T = simulation time-scan interval (sec) K = sensitivity parameter used in modified Pitt car-following model sl( x ) = position of lead vehicle at time x as measured from upstream (ft) R = perception-reaction time (assumed to be equal for all vehicles) (sec) sf( x ) = position of follower vehicle at time x as measured from upstream (ft) Ll = length of lead vehicle plus a buffer based on jam density (ft) h = time headway parameter in Pitt car-following model (buffer headway) (sec) vf( x ) = speed of follower vehicle at time x (ft/s) vl( x ) = speed of lead vehicle at time x (ft/s) al( x ) = acceleration of l ead vehicle at time x (ft/s2) afe = acceleration of follower vehicle as computed from safe following (ft/s2) vlmin = minimum speed of lead vehicle (ft/s) almin( vf) = minimum acceleration of lead vehi cle (maximum deceleration) (ft/s2) For purposes of the research, the initial move ment of the vehicles was defined once they reached the speed of one foot per second. Furthermore, it was found that an appropriate value for the parameter K should be greater than 1. This is taking into consideration the assumption that a driver in a queue behaves significantly different than a driv er in free-flowing traffic. The best fit for this value in the study was 1.25. Greenshields [12] indicates that the vehicle head ways become fairly constant after the fifth vehicle. This indicates that the first four cars should not be used for the discharge headway calculations. These cars are the major carriers of the start-up lost time. The results of this research include only passenger cars. In the study it was found, as expected, that longe r vehicle length results in longer discharge headways. Furthermore, it was found that the ex pansion wave speed is independent of the Free

PAGE 25

25 Flow Speed (FFS) but it increases when the follo wing distances are closer The expansion wave may slow down when the vehicles average sp eed reaches 30 mph. This is reasonable since acceleration decreases at higher speeds. This research indicates that the impact of a tr uck is greater if the truck is positioned in the first few positions. Additionally, it is recommended that the K parameter should be calibrated with vehicles in the second and third position since they are the vehicles most affected. Bonneson [13] studied and summarized prev ious researches of modeling discharge headways at signalized intersec tions. Bonneson selected a model based on vehicle and driver capabilities, including driver r eaction time, driver acceleration, a nd vehicle speed. Then, he used field data from five different sites to calibrate the model. The Briggs headway model [14] assumes that the acceleration for each vehicle in queue remains constant. This model separates the vehi cles not by position but by distance to the stop bar in comparison with the distance they need to reach their desired speed. This distance is denoted as d in Equation 2-5. If maxd d n [2-5] Then A n d A n d T hn) 1 ( 2 2 [2-6] Otherwise q nV d T h [2-7] With 2 max2qV d a [2-8]

PAGE 26

26 Where: n = queue position ( n = 1, 2, 3) d = distance between vehicles in a stopped queue (ft) dmax = distance traveled to reach speed Vq (ft) hn = headway of the nth queued vehicle (sec) T = driver starting response time (sec) a = constant acceleration of queued vehicles Vq = desired speed of queued traffic (ft/s) Messer and Fambro [15] found that regardless of the position of the driver, the response time was 1.0 sec. Although it was found that an additional delay of 2.0 sec should be allocated to the first driver. They also found that an average length for each queue position is 25 ft. To these values it is important to add the addi tional length and delay due the vehicle type. In this report the vehicle compositi on of the queue is not specified. Buhr et al. [16] conducted a study of driver ac celeration charact eristics on freeway ramps. This study considers only passenger cars from a stopped condition. The authors determined that acceleration decreased linearly with incr easing speed according to Equation 2-9. max max1 V aA V [2-9] Where: a = instantaneous acceleration (ft/s2) Amax = maximum acceleration (ft/s2) V = velocity of vehicle (ft/s) Vmax = maximum speed corresponding to zero acceleration (ft/s) For on-ramps at level terrain, the authors found an Amax = 15 ft/s2and Vmax = 60 ft/s. A later study conducted by Evans and Rothery [17], studied the acceleration and sp eed characteristics of queued drivers. The results show an exponent ially increasing speed to a desired speed of approximately 50 ft/s. This speed is reached by the vehicle in 25th position, which can be considered as normal cruising sp eed when crossing the stop bar.

PAGE 27

27 Previous studies indicate that the discharge headway ( h ) between the nth vehicle and the ( n-1) th vehicle has two components. First, the time that it takes from the beginning of the green time to the driver to start movi ng. This can be estimated as + n T where is the additional time for the first driver and T is the individual response time. Th e second part is the time that it takes the vehicle to reac h the stop bar. Briggs assumed that each vehicle in queue will occupy the same amount of space. The final form of the proposed headway model is shown in Equation 2-10. ()(1) 1 maxmax stnstn nVV d hN Va [2-10] Where: hn = headway of the nth queued vehicles (sec) = additional response time of the first queued driver (sec) 11 if it is the first vehicle 0 otherwise N T = driver starting response time (sec) d = distance between vehicles in a stopped queue (ft) Vmax = maximum speed (ft/s) Vst( n ) = stop line speed of th e nth queued vehicle (ft/s) Amax = maximum acceleration (ft/s2) This approach does not consid er any kind of mixed traffic. Another limitation in this model is that it does not consid er any variation in the inter-ve hicle length or any other buffer length between the two vehicles. In conclusion, this study says that the mini mum discharge headway is dependent on driver response time, desired speed and traffic pressure for each movement. It states that the minimum discharge headway is not reached until the eighth vehicle. This model also suggests that under ideal conditions the discharge headway should be shorter than 2.0 sec/veh. Akelik et al. [19] describe an exponential queue discharge flow and sp eed model. In this study they model queue discharge speed in additi on to the headway. By including speed it is

PAGE 28

28 easier to develop relationships for traffic para meters like vehicle spacing, density, time and space occupancy ratios, gap time, occupancy time, space time and acceleration characteristics. Equations 2-11, 2-12, and 2-13 show the models for speed, flow and headway developed by Akelik, et al [19]: ] 1 [) ( tr t m n sve v v [2-11] ] 1 [) ( tr t m n sqe q q [2-12] ] 1 [) ( tr t m n sqe h h [2-13] Where: vs = queue discharge speed at time t (km/h) vn = maximum queue discharge speed (km/h) mv = parameter t = time since the start of green (seconds) tr = start response time (constant) average from all drivers in 1st position qs = queue discharge flow rate at time t (veh/h) qn = maximum queue discharge flow rate (veh/h) mq = parameter hs = queue discharge headway at time t (seconds) hs = 3600/ qs hn = minimum queue discharge headway (seconds) hq =3600/ qn Pipes [20] describes the ideal space or dist ance headway as one car length for every ten miles per hour of speed at which the follower ve hicle is traveling. Th e resulting equation is shown in Equation 2-14. n n n MIN n nL t v L t x t x d ) 10 )( 47 1 ( ) ( ) ( ) (1 1 min [2-14] Where: dmin = distance between vehicles at time t (ft) xn = location of lead vehicle at time t (ft) xn+1 = location of follower vehicle at time t (ft) Ln = length of lead vehicle (ft) vn+1 = speed of the following vehicle at time t (mph)

PAGE 29

29 In this case the headway depends more on the leng th of the lead vehicle. For example, say a passenger car is following a truck, the ideal distan ce will be using the lead vehicles length. In this case it will result in an extremely large h eadway. On the other hand, if a truck is following a passenger car, the model gives a smaller headway. It is obvious this model is not taking in consideration the driver behavior and the braking capabilities of the vehicles. Previous studies have found that it takes longer to a truck to stop. Therefore, the model is not compatible with real life scenarios. This model is relatively easy to use, the mathematical operations are simple. The big limitation of this model is that it does not take in to consideration any kind of interaction between vehicles. The acceleration is considered to be constant and there is no way to allocate heavy vehicles. As appear in May [21], Forbes theory a pproaches the car-following by considering the reaction time of the follower. The minimum gap between the vehicles should be greater than the reaction time of the drivers. In addition to th e minimum gap, this model considered the length of the vehicles. In Equation 2-15 is expr essed this mathematical relationship. ) ( t v L t hn n MIN [2-15] Where: h = time headway (seconds) t = reaction time, assumed to be 1.5 sec (seconds) Ln = length of lead vehicle (ft) vn = speed of the lead vehicle at time t (mph) May [21] also presents the studies devel oped by a group of researchers from General Motors about car-following theories These studies were more extensive than the studies made for Pipes model and Forbes model. It also ha s a particular importance because it is based on an empirical model. After four previous in tents or versions, GM came up with the 5th and final

PAGE 30

30 model. This eliminates the discontinuities in the previous versions. This final model is shown in Equation 2-16. ) ( ) ( ) ( ) ( ) ( ) (1 1 1 1t v t v t x t x t t v t t an n l n n m n m l n [2-16] Where: an+1( x ) = acceleration of follo wer vehicle at time x (ft/s2) t = reaction time (seconds) = sensitivity para meter with speed a nd distance exponents m and l respectively m = speed exponent l = distance exponent vn+ 1( x ) = speed of follower vehicle at time x (ft/s) vn( x ) = speed of lead vehicle at time x (ft/s) xn( x ) = position of lead vehicle at time x as measured from upstream (ft) xn+1( x ) = position of follower vehicle at time x as measured from upstream (ft) This model includes the accelera tion characteristics of the fo llowing vehicle and yet it does not include a variation for vehicle type or size. However, the parameters m and l can be calibrated by individual vehicle types to adju st the acceleration of the following vehicle. Another limitation of this model is the lack of consideration for the vehicles length. Long [22] found that passenger cars, SUVs, and vans length averaged around 15 ft. These vehicles length vary from 10.9 to 19.7 ft, with approximately two-thirds of the sample range between 13 and 16 ft. It was also found that when any of these cars were pulling a trailer, their total length increase in average to 35 ft. The truck distribution was not as close to a normal distribution as the PCs were. Additionally, only an approximated 12 % of the trucks were as l ong as the WB-50 design length (55 ft) or shorter. Long [22] found that combination trucks typi cally have a length of 65 ft. Using the grouping that AASHTO has follow for years regarding acceleration capabilities, typical lengths of 15, 65 and 30 ft can be used for PCs, combinations trucks and all others

PAGE 31

31 vehicle respectively. This research also specifie s some other types of truc ks that can potentially be used as guidance. Then an expected average vehicle length (EVL ) could be estimated w ith weighted average using the expected proportion for each one of the groups. Long also recommend a reasonable inter vehicle spacing of 12 ft in contrast to the 3 ft default value in CORSIM. Summary of HCM 2000 PCE Guidelines Currently, a PCE value of 2.0 is specified for all large vehicles. No distinction is made between trucks, recreat ional vehicles, and buses in the calculation of the adjusted saturation flow rate at signa lized intersections. It applies to any vehicle with more four tires in cont act with the driving surface. The headways of the first four to six vehicles in queue are no t considered as part of the saturation headway. A default start-up lost time (S LT) of 2.0 seconds, if field measurements are not available, is recommended. Summary of Previous PCE Studies There are no recent studies that have examin ed or determined heavy vehicle PCE values for signalized intersections. There are previous studies that demonstr ate how heavy vehicles have an impact on traffic streams at an intersec tion. Of all the research reviewed, headway was by far the most common performance measures used to base PCE values on. Significant findings from these studies are as follows: Molina derived a different expression for the additional effect of a heavy vehicle in the first position of a queue. Molina also f ound that position in queue did not have a pronounced effect on twoand threeaxle and single-unit trucks. Benekohal and Zhao concluded that other than being in the first pos ition, the position of the heavy vehicle in the rest of the queu e does not matter, but yet the number of cars behind the heavy vehicle has a significant impact. Benekohal and Zhao concluded that at signaliz ed intersections, the additional delay caused by larger trucks should be used for calcul ating PCEs. They introduced a new term, DPCE, meaning a delay-base d calculation for PCE.

PAGE 32

32 Kockleman and Shabih found that three factor s influence the vehicl e headways: length, performance and driver behavi or. In addition they recomm ended PCE values of 1.07, 1.41, 1.34, and 1.14 for small SUV, long SUV, vans and pick-up trucks respectively. Bonneson et al. developed new and revised adju stments to the saturation flow equation. They estimated a PCE value 1.74 for heavy vehicl es, but that was based on a data set with minimal truck observations. Perez-Cartagena and Tarko stated that the fHV is adequate for use in Indiana, although they recommend an additional factor for population. Li and Prevedouros found that the mini mum headway is not reached until the 9th vehicle crosses the stop bar. They also recomme nded a mean start-up r eaction time of 1.76 seconds with a standard deviation of 0.61. Summary of Car-Following Models There were several car-following models reviewed in this chapter including from the first few models developed to the most recent ones. Table 2-1 summarizes the details that were taken into consideration at the moment of selecti ng which car-following mode l was going to be used. Table 2-1. Car-following models comparison Pipes Forbes GM Modified Pitt Constant acceleration Yes Yes No No Acceleration of the leading vehicle No No No Yes Speed at stop bar No No Yes Yes Following distance Fixed thr ough vehicle type Fixed Fixed Fixed but can be adjusted with vehicles length Vehicle length Yes Yes No Yes Length variation for vehicle type No but can be adjusted No but can be adjusted No Yes but have to be adjusted Calibration difficulty High High High Medium Implementation difficulty Low Low Low Low Computational efficiency High High Medium Medium Of these five car-following models studied, the modified Pitt model fits better to apply the model to a signalized intersection with heavy vehicles in the stream. In the model, the summation of the length of each vehicle plus a bu ffer length determined at jam density is known as L For passenger cars, a value of 20 ft is used for this length L This value should be reconsidered due the findings in Longs study [22]. This value should be the sum of two parts.

PAGE 33

33 The first part should be the lengt h of the leading vehicle, and second the minimum or allowable inter-vehicle spacing. By doing this, a more de tailed composition of the traffic stream can be reached.

PAGE 34

34 CHAPTER 3 RESEARCH APPROACH This chapter describes the research approach that was used to accomplish the objectives of this study. More specifically, it will discuss the methodological ap proach, field data collection, simulation model development, and the simulation experiments. Methodological Approach As discussed in the literatur e review chapter, two differe nt methodological approaches have been used in previous PCE related studies One was based on the concept of a delay-based passenger car equivalent for tr ucks. The other was based on the concept of a time headway passenger car equivalent. The latt er is the approach that is currently used in the Highway Capacity Manual [1]. One of the main objectives of this project was to update the PCE values used for large trucks in FDOTs ARTPLAN so ftware [23]. Since the ARTPLAN calculation methodologies are largely based on the HCM analys is framework, it was decided to use the time headway approach for purposes of consistency. In the headway based approach to PCE calcula tion, it is important to recognize that the time headway value between successive vehicles is a function of both the leading vehicle and the trailing vehicle. Thus, a PCE value that is dete rmined for a particular type of vehicle must account for a variety of different vehicles that may precede it in a queue. For example, the headway between a passenger car (l eader) and large truck (follower) will be differe nt than that between a large truck and a large truck, all else be ing equal. Of course, it is not reasonable to expect practitioners to collect the exact sequence of vehicles in queue during data collection activities. Generally, the data collection is limited to just a percentage of trucks in the traffic stream, not their actual positions in the traffic stream as well. Furthermore, this sequence will usually be different from cycle-to-cycle. T hus, the PCE values that are derived must be

PAGE 35

35 somewhat generalized so that the sequence of vehicles is not an input to the selection of a PCE value. While it might seem desirable to develop th e PCE values based on a large number of the different leader-follower vehicle combinations, there are some practical concerns with this approach. First, certain combinations of vehicles will occur very rarely in the traffic stream; for example, a recreational vehicle pulli ng a trailer that is following a motorcycle, or vice versa, as the individual frequencies of o ccurrence for these vehicle types is quite small. Second, from a data collection perspective again, it is only reasonable to expect practitioners to classify vehicles into just a few different categories. Therefore, for this project, it was decided to consider just three different truck types, in a ddition to the passenger vehicle. Different types of trucks have different im pacts on the traffic stream; for example, one single-unit truck does not take up the same amount of space as one tractor+semi-trailer, and their acceleration capabilities are most likely different as well. Therefore, heavy vehicles were categorized as either a small, medium or large truck depe nding on their size and operational characteristics. The details of each category will be described in a later section in this chapter. The end result of this project is the development of three different PCE values applicable to three categories of heavy vehi cles which can be used to calculate a heavy vehicle factor ( fHV) analogous to the current HCM meth od, as shown in Equation 3-1. 1 1(1)(1)(1)HV STSTMTMTLTLTf PEPEPE [3-1] Where: fHV = adjustment factor for heavy vehicles in traffic stream Pi = proportion of truck type i in traffic stream Ei = PCE factor for truck type i i = LT for large truck, MT for medium truck, and ST for small truck

PAGE 36

36 According to the procedure in the HCM to obtain the saturati on flow rate, it was necessary to measure the saturation headway at each inters ection with at least eight vehicles in queue. Since it was established that the headway of each vehicle depends on the leading and trailing vehicle and that there were thr ee different truck types under study, the number of possible leaderfollower combinations is 16 (42). Table 3-1 enumerates these 16 combinations. Table 3-1. Possible pairing combin ations for four vehicle types* PC PC PC ST PC MT PC LT ST PC ST ST ST MT ST LT MT PC MT ST MT MT MT LT LT PC LT ST LT MT LT LT PC = passenger car, ST = small truck, MT = medium truck, LT = large truck Since the focus of this study was on truck PC E values, it was obviously necessary to find sites with a relatively high percentage of large trucks. For sites with less than 10% trucks in the traffic stream, a very large percentage of the que ues would have only zero or one truck in it. Thus, the number of cycles that would need to be collected to obtain a significant number of queues with multiple trucks present would be ex treme, thus leading to a very lengthy and inefficient data collection process. Thus, only sites with a truck percentage of at least 10% were considered for data collection. However, sites w ith persistent queues of at least 8 vehicles and 10% or more large trucks are not very common. Sites with a lot of traffic (thus generating the necessary queue lengths) typically have small truck percentages, and sites with high truck percentages (such as along truc k routes) do not typically have high overall volumes. With the data collection resources available for this project (i.e., time, money, labor), it was not feasible to obtain enough data from enough sites to facilitate the objectives of this project from field data alone. Therefore, the decision was made to collect as much field data as project resources would allow, and then use these field data to calibrate a simulation model that w ould be used to provide the full data set upon which to base the developm ent of the truck PCEs. The collection of the

PAGE 37

37 field data is described in the ne xt section, and the development and application of the simulation model in the section after that. Field Data Collection In order to reduce the impact of additional factors that could a ffect the effectiveness of the intersection, and to obtain the n ecessary amount of data, it was esse ntial for the site s selected to contain certain characteristics. The criteria to select these si tes for field data collection are outlined below. Site Selection Criteria Geometry Typical four leg intersection (tur ning radii at or close to 90) At least one site with only one through lane The other sites shoul d have two or three travel lanes in the through direction The approach must have an exclusive left turn lane(s), but the queue cannot spill back onto through lanes Sites with an exclusive right turn lane are preferred, but not absolutely necessary if the site has a small percentage of right turning vehicles Level terrain is preferred, but site s with small grades are acceptable No curbside parking or bus stops near the inte rsection, or other external factors that will significantly influence th e saturation flow rate Traffic Queue lengths of at least 10 veh/lane at the beginning of green shoul d be regularly present during the data collection period. This is a function of the signal g/C ratio and cycle length, in addition to the overall traffic demand. For example, for a g/C ratio of 0.4 and a cycle length of 60 seconds, an average hourly flow rate of at least 1000 veh/hr/lane is needed. Additionally, if progression is favorable, then the volumes would need to be adjusted upward slightly, as a higher percenta ge of vehicles would be arriving on green. A heavy vehicle percentage of 10% or higher in the traffic stream.

PAGE 38

38 The operations of the observed approach must not be impacted by a downstream queue. That is, vehicles must be able to de part freely from th e subject approach. From these criteria, and assistance from FDOT personnel, six sites were ultimately chosen for field data collection. These sites are listed in Table 3-2. Table 3-2. Data collection cites Site # Intersection street names City 1 SW Williston Rd / SW 34th St Gainesville 2 Waldo Rd / Univers ity Ave Gainesville 3 US 41 / SR 50 Brooksville 4 US 301 / SR 50 Brooksville 5 SR 326 / CR 200A Ocala 6 John Young Parkway / Colonial Drive Orlando Sites Selected Herein are shown aerials and ground le vel views from each site selected. Figure 3-1. SW Williston Rd / SW 34th St aerial view (eastbound on Williston Rd) North

PAGE 39

39 Figure 3-2. SW Williston Rd / SW 34th St ground level view (eastbound on Williston Rd) Figure 3-3. University Ave. / Wal do Rd aerial view (northbound on Waldo Rd) North

PAGE 40

40 Figure 3-4. University Ave. / Waldo Rd ground level view (northbound on Waldo Rd) Figure 3-5. US 41 / SR 50 aerial view (westbound on SR50) North

PAGE 41

41 Figure 3-6. US 41 / SR 50 gr ound level view (westbound on SR50) Figure 3-7. US 301 / SR 50 aer ial view (eastbound on SR 50) North

PAGE 42

42 Figure 3-8. US 301 / SR 50 ground level view (eastbound on SR 50) Figure 3-9. CR 326 / SR 200A ground level view (westbound on CR 326)

PAGE 43

43 Figure 3-10. John Young Parkway / Colonial Dr aerial view (southbound on John Young Pkwy) Figure 3-11. John Young Pkwy / Colonial Dr ground level view (southbound on John Young Pkwy) North

PAGE 44

44 Data Collection Methods Two different methods for data collection we re used. One method used a single camera along with equipment installed in the signal controller cabinet to obtain signal status information concurrent with the video signal. This method is referred to as method 1, and was used for only the Williston Rd/34th St site. The other method used tw o cameras to obtain both traffic and signal status information, and is referred to as method 2. Each of these methods is described in the following sections. Data collection equipment for method 1 For this method, it was necessary to be able to set up the following devices inside of the signal controller cabinet: a VCR, a signal encodi ng device, and current sensors that attach (passively) to each of the green bulb/LED power wires. Figure A-1 illustrates the necessary connections. A picture of the installation of the equipment in a signal cont roller cabinet is shown in Figure A-2. A video camera was mounted on the mast arm facing the approach of interest. An additional constraint for this setup was that the camera be mounted on a mast arm that is in the same quadrant as the signal controller cabinet to easily facilitate runni ng the video/power cable from the video camera to the controller cabinet. Figure 3-12 illustrates the camera setup at the Williston Rd/34th St intersection. The stop bar and back of queue must be visibl e within the camera field-of-view (FOV), as demonstrated in Figure 3-13. Figure 3-13 also shows how the vi deo image and the signal status information are combined into a composite view to facilitate data reduction. This is accomplished through special processing hardware and software that is too detailed to explain here. For more information on this system, see Washburn et al. [2].

PAGE 45

45 Figure 3-12. Preferred video camera m ounting location (plan view) for method 1 Figure 3-13. Screen capture (low resolution) of video image from Williston Rd/34th Street site in Gainesville (method 1) Camera FOV

PAGE 46

46 Data collection equipment for method 2 To be able to implement the first method for data collection requires the cooperation of the local agency that maintains the signals. Du e to an existing relationship with the City of Gainesville, this cooperation was easily facilitated However, it was not so easily facilitated in the other jurisdictions. Thus, it was decided to use a different da ta collection approach for sites outside of Gainesville to avoid this complicati on. This second method requires no access to the control cabinet or the need of th e presence of local agency staff. This method consists of having two cameras at the site at ground level. One cam era is placed in a position where it can focus on the traffic signal head and at the same time see a fair amount of the queue. The second camera is placed where it has a FOV that includes the stop bar and the first couple of cars in the queue. This setup is illustrated in Figure 3-14. Figure 3-14. Camera setup fo r data collection (method 2) Green light FOV Stop bar FOV

PAGE 47

47 For this method, if the back of queue was not visible within the camera FOV for some cycles, it was recorded manually. Method 2 was used for data collec tion at the other five sites. While one of the other sites was in Gainesville, method 1 could not be used at this site due to complications of running the necessary cables from the vide o camera through the in-ground conduit to the signal controller cabinet. To facilitate data reduction w ith this method, the two camera views had to be combined into a single composite image, along with a timer, as shown in Figure 3-15. This was accomplished with a video mixer, a software program that generated the timer, and another software program that merged the timer with the output of the two camera images from the mixer. Figure 3-15. Screen capture of the composite video image (method 2) Timer in seconds Traffic Signal Heads Stop bar Overlapped image

PAGE 48

48 Data Collection Periods For the data collection performed at Williston Rd/34th St site, there was more flexibility in how much data could be collected, and at what tim es data could be collected. This was because with method 1, the VCR in the controller cabinet could be prog rammed to record any time period, and the City of Gainesville allowed the research team to access the cabinet to change VCR tapes at our discretion. For data collection with method 2, it was necessary to be present on site during the entire duration of the data collection. Thus, due to this issue and travel costs and constraints, data collection was limited to a 4-hr period at the sites outside of Gainesville. The data collection periods for the Williston Rd/34th St site (using method 1) are shown in Table 3-3. Table 3-3. Data collection periods for Williston Rd/34th Street site (method 1) Tape # Date Time(s) 2 3/2/2006 6:30 9:30 am; 11:30 am 1:30 pm; 3:30 6:30 pm 5 3/31/2006 8:30 am 4:30 pm 6 4/12/2006 6:30 am 2:30 pm The data collection periods for the sites using method 2 are shown in Table 3-4. Table 3-4. Data collection pe riods for other sites (method 2) Site City View1 Lanes Tape num Time interval Total time hr:min Date 1 Gainesville Stop bar 2 1 3:30 5:30 pm 2 6/26/06 2 Ocala Stop bar 1 1 8:00 12:00 pm 4 6/29/06 Green 2 8:00 12:00 pm 4 6/29/06 3 Orlando Stop bar 3 1 8:00 12:00 pm 4 6/30/06 Green 2 8:00 12:00 pm 4 6/30/06 1 Gainesville Stop bar 2 1 2:00 6:00 pm 4 7/05/06 Green 2 2:00 6:00 pm 4 7/05/06 4 Brooksville Stop bar 2 1 8:00 12:00 pm 4 7/07/06 Green 2 8:00 12:00 pm 4 7/07/06 5 Brooksville Stop bar 1 1 9: 05 11:45 pm 2:40 7/07/06 Green 2 9:35 11:45 pm 2:10 7/0706 1 This indicates the field-of-view perspective of the cameras used for data collection at the site.

PAGE 49

49 Data Reduction The composite image videos were manually proc essed. The information that was recorded from each video included the time when the signal turned green and the time when the front axle of each vehicle in the queue crossed the stop bar. Also recorded was the type of each vehicle in queue, according to a predetermined vehicle clas sification scheme (either passenger car, small truck, medium truck, or large truc k). Examples of the types of vehi cles that fall into each of the truck categories are shown in Appendix B The indi vidual headway data and lost times were used to develop the truck passenger ca r equivalent values for the thr ee different classifications of trucks, as discussed in Chapter 4. Due to equipment problems, tapes 2 and 6 at the Williston Rd/34th St site unfortunately did not have the start of green information. Howeve r, all other information could still be obtained from these videos. It should be noted that the precision of the timing measurements is limited to 0.0333 seconds. This is a re sult of using video cameras and reco rding equipment th at utilize the standard frame rate of 30 frames per second. Simulation Modeling This section describes the development of th e simulation program, which was used for the generation of a much larger data set to base the development of the PCE values upon. While there are a variety of commercial simulation progr ams on the market that can simulate traffic flow at signalized intersections, it was decided to develop a custom simulation program. This decision was made for several reasons: 1) ma ny commercial software programs do not readily provide detailed documentation about the unde rlying car movement models; 2) the car movement models contained in some programs are not very robust, particul arly with respect to heavy vehicles, or do not provide for user modi fication of some key parameters within those

PAGE 50

50 models; and 3) with a custom si mulation program, custom preand post-processing routines can also be developed for increased efficiency. Thus, by developing a custom simulation program, all aspects of the car movement models and othe r features of the program can be completely controlled. The development of the simulation program involved aspects such as creating a user interface for specification of simu lation scenarios and run settings and implementation of a carfollowing model. These efforts are de scribed in the following sections. Car-Following Model The foundation of a traffic simulation program is the underlying mathematical models that describe the movement of vehicles along the ro adway system. Thus, th e first task in the development of the simulation program was the selection of a car-movement model that was suitable to a queue discharge situa tion at a signalized intersection. Based upon the literature review in Chapter 2 for car-followi ng models, and the needs of this research project, the Modified Pitt model was selected for implementation [11]. The Modified Pitt car-following model calculates an acceleration value for the trailing vehicle based on intuitive parameters, such as the speed and acceleration of the lead vehicle, the speed of the trailing vehicle, the relative positi on of the lead and trail vehicles, as well as a desired headway. Car-following models are gene rally based on a driving rule, such as a desired following distance or following headway. The Modified Pitt model is based on the rule of a desired following headway. As indicated before, the headway of each ve hicle depends of the leading and trailing vehicle and this model takes into considerati on the physical and operati onal characteristics of both. The main form of the model is shown again in Equation 3-2.

PAGE 51

51 T h T T R t a T R t v R t v v h L R t s R t s K T t al l f f l f l f2 1 2 12 [3-2] Where: af( t +T) = acceleration of fo llower vehicle at time t + T in ft/s2 al( t + R ) = acceleration of l ead vehicle at time t + R in ft/s2 sl( t + R ) = position of lead vehicle at time t + R as measured from upstream, in ft sf( t + R ) = position of follower vehicle at time t + R as measured from upstream, in ft vf( t + R ) = speed of follower vehicle at time t + R in ft/s vl( t + R ) = speed of lead vehicle at time t + R in ft/s Ll = length of lead vehicle plus a buffer based on jam density, in ft h = time headway parameter (refers to headway between rear bumper plus a buffer of lead vehicle to front bumper of follower), in seconds T = simulation time-scan interval, in seconds T = current simulation time step, in seconds R = perception-reaction time, in seconds K = sensitivity parameter (unit less) For application to this project, the value of the L parameter varied based on one of the four different vehicle types. The time headway parameter ( h ) was set up as a random variable, rather than a constant value, to introdu ce an additional stochastic elemen t to the model. Its value was based on a normal distribution to represent the more realistic scenario that desired headways vary by driver. The mean and standard deviation for this distribution could be specified for each of the four vehicle types. Thus, desired headwa ys can vary by driver, as well as by vehicle category. Additional details on the Modifi ed Pitt car-following model can be found in Cohen [24]. Program Development The simulation program was written in Visual Basi c 6. A screen capture of the main user interface is shown in Figure 3-16.

PAGE 52

52 Figure 3-16. Simulation program user interface The top of the user interface is where all of the vehicle characteristics and model parameters are specified. The lower part of the user interface is where the simulation run options were specified. Four options can be specified: 1. Generate a queue of vehicles randomly, ba sed on the specified vehi cle proportions (the queue length is a constant 8 vehicles). 2. Generate a specific number of each vehi cle type in the queue, in random positions. 3. Generate a specific vehicle type in a speci fic queue position, for each of 8 total queue positions. 4. Reading any number of pr e-specified queue configurations from an input file.

PAGE 53

53 The first three options were used primarily for model testing purposes. The fourth option was used to facilitate the running of a larg e number of pre-specified simulation scenarios according to the experimental design, described in a later section. The program includes a traffic animation co mponent. This component animates the vehicle trajectory information recorded during the simulation process. It updates the position, speed, and acceleration of each ve hicle in the queue every tenth of a second. This screen includes the signal status and the elapsed time fr om the beginning of the green. Additionally, it includes a table that displays key vehicle prope rties during each time step of the animation, which is used primarily for diagnostic purposes. Screen captures of this animation screen are shown in Appendix C. Experimental Design As mentioned previously, the main purpose of the field data was for calibration of the simulation program. The simulation program was us ed to generate the data for the development of the passenger car equivalent va lues for trucks, as well as revi sed lost time estimates. One obvious limitation with the field data is that a limited number of vehicle type-queue position combinations were observed, and of the combina tions that were observed, some were observed a very limited number of times, some only once. With the use of simulati on, a wide variety queue combinations (i.e., different vehicle types in different queue positions) can be generated, as well as any number of replications of a specific queue combination. One of the first decisions to make regarding the experimental design is which variables to include. The study on saturation flow rate by Bonnes on [8] identified severa l factors that affect this value, such as speed limit, number of lanes, and traffic pressure. As discussed in Chapter 4, none of these factors were found to be significant, or at least were inconclusive, in the field data. Thus, none of these factors were included in the experimental desi gn. Some of these variables

PAGE 54

54 are still inputs to the simulation program, such as speed limit, but these were fixed at one average value. Furthermore, given the anticipate d size of the simulation analysis data set, it was felt that simulation and analysis resources would be better spent focusing on just the truckspecific aspects of the queue discharge mechanism, which was the primary concern for this project. With this approach, any revised PCE va lues resulting from this study could be used in combination with the other factors developed as part of the Bonnes on study. While the Bonneson study did identify a sing le revised truck PCE value, the authors admit that the examination of the effects of trucks on capacity was not a variable of primary concern; thus, truck percentages were quite low in the field data th ey collected for their proj ect. In fact, one of the specific recommendations from that study was to further investig ate truck PCE values. Therefore, the experimental design consisted of just varying the vehicle type by queue position, for a fixed queue length, and leaving other factors fixed at representative values. Note that certain factors in the simulation program, such as ma ximum acceleration, are varied randomly by vehicle/driver for each queue genera tion, according to a mean value and standard deviation. However, these factor s are part of the stoc hastic simulation proce ss, and not factors to be estimated as part of the analysis process. In working within the framework of the HCM guidance for measuring lost time and saturation flow rate, the queue length for the experimental de sign was set at eight vehicles for each combination. Thus, with the four differe nt vehicle types, and a queue length of eight vehicles, the number of possi ble combinations is 65,536 (48). In order to reduce the computational burden due to such a large number of combinations, and to better reflect reality with the specific combinati ons, the number of combinati ons was reduced. Among the 65,536 possible combinations for four vehicle types an d eight queue positions are many combinations

PAGE 55

55 that include a high percentage of trucks in the queue. Since qu eues with a high percentage of trucks were extremely rare in the field, it was decided to eliminate all queue combinations that consisted of a truck percentage of more than 50%. Of all the field data queues, very few queues had more than three trucks (out of a queue length of eight vehicl es), and only one queue had five trucks. No queues had more than 5 trucks (out of eight). Elimin ation of all queue combinations with more than four trucks resulted in a tota l of 7,459 combinations, which comprised the final experimental design.

PAGE 56

56 CHAPTER 4 ANALYSIS OF FIELD DATA AND CALIBRATION OF SIMULATION MODEL This chapter describes the reduction and anal ysis of the field data, as well as the development of a model to fit the traffic flow data obtained from the field. It also describes in detail the calibration of the simulation program using the collected field data Summary of Data Reduction From the reduced field data, only signal cycles that had queues with at least 8 vehicles were retained for data analysis. A summary of this data set is include d in Appendix D. This table shows the different queue compositions obs erved in the field along with their respective frequencies. Note that for not ational convenience, vehicle type s are expressed in a numbered format, where 1 is a passenger car, 2 is a small tr uck, 3 is a medium truck and 4 is a large truck. This data set consisted of a total of 403 cycl es, where 174 cycles were passenger cars only, 126 cycles had 1 truck, 68 cycles had 2 trucks, 28 cycles had 3 trucks and 7 cycles had 4 trucks in queue. Some of these queue compositions per cycle were repeated; therefore there were a total of 110 different observations. In addition, only one queue was observed to have 5 trucks but it was left out of the calibration process sin ce the experimental desi gn was constrained to no more than 4 trucks in queue. This will be explained in more detail later in this chapter. In chapter 3 it was discussed that the time h eadway value between successive vehicles is a function of both the leading vehi cle and the trailing ve hicle. A summary of the average time headway for the 16 possible vehicle pairs (for 4 vehi cle types) is shown in Appendix E. In Table 4-1 is a summary of those tables for vehicles in queue positions 2-8 and 5-8. Headways of vehicles in position 1 are not included in these tables since only 231 out of 403 cycles included the start of green.

PAGE 57

57 Table 4-1. Average headway and frequenc ies for each leader-follower combination Average headways of vehicle pairs in positions 2 through 8 Trail vehicle1 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Lead vehicle1 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 Frequency 417 51 37 61 45 8 5 7 36 4 4 1 65 6 7 16 Mean 2.40 3.13 3.34 4.70 3.01 3.85 4.92 5.61 3.67 4.86 5.70 4.24 4.14 4.42 4.97 5.09 Average headways of vehicle pairs in positions 5 through 8 Trail vehicle1 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Lead vehicle1 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 Frequency 232 29 21 38 28 3 4 3 21 2 3 1 39 2 3 11 Mean 2.19 2.82 2.72 4.13 2.86 3.08 4.88 4.22 3.74 5.61 4.59 4.24 4.13 4.50 4.46 5.23 1 Vehicle Types: 1 = passenger ca r, 2 = small truck, 3 = medi um truck, 4 = large truck Time Headways Time headways are often used to estimate the im pact of trucks in the traffic stream, but to specifically calculate the saturation flow rate it is necessary to use the average headway of vehicles in positions 5 through 8 ( 8 5 h) or also known as the saturation headway (hSAT). It should be noted that this definition is consistent with the HCM, alt hough some studies have found that the saturation headway is not achie ved until the sixth or later vehicle in the queue. The concept of the saturation headway, consistent with the HC M definition, is illustrated mathematically with the following equations. 1. Saturation headway (i.e., the average h eadway for vehicles in positions 5-8) 44 8T T hSAT [4-1] 2. Average headway for vehicles in positions 1-8 88 8 1T h [4-2] 3. Average headway for vehicles in positions 2-8 71 8 8 2T T h [4-3] Where: Ti = the time it takes for vehicle i to cross the stop bar Table 4-2 summarizes the headways calculated from the field data.

PAGE 58

58 Table 4-2. Time headways from field data All vehicles All vehicles with start-up time Pa ssenger cars only Passenger cars only with start-up time hSAT 2.87 2.83 2.18 2.23 h1-8 N/A 2.98 N/A 2.48 h2-8 3.11 N/A 2.36 N/A Several models were specified in the statis tical software package STATISTICA [25] to attempt to reasonably describe th e impacts of trucks on queue di scharge for these field data. These models considered queue composition, truck percentages, and site characteristics. The specifications were done using a non-linear regression analysis with a confidence level of 95%. A summary of this model development is described in the following paragraphs. Model 1 The first model (Equation 4-4) us ed an alternate form of the fHV equation (HCM chapter 16) to estimate the average saturated headway ( hSAT). This analysis used the data that did not contain start-up reaction time values. 1 1 1 13 2 1 b Pct b Pct b Pct h hLT MT ST SPC SAT [4-4] Where: hSAT = saturation headway hSPC = saturation headway of passenger cars in a passenge r car only queue PctST = percentage of small trucks in queue PctMT = percentage of medium trucks in queue PctLT = percentage of large trucks in queue In this model, the coefficients b1, b2 and b3 represent the ET values for small, medium and large trucks respectively. The estimated coeffi cient values are shown in Table 4-3. An R2 value of 0.3591 was obtained for this model. Table 4-3. Results from STATISTICA for model 1 Coefficients Estimate Standard error t-value hSPC 2.216044 0.033277 66.59442 b1 1.655451 0.165439 10.00640 b2 1.771816 0.196516 9.01615 b3 2.677482 0.132386 20.22480

PAGE 59

59 Model 2 The second model (Equation 4-5) is the same as Model 1, but this model estimates the average headway for vehicles in position 2 through 8, where hPC is the average headway for passenger cars. Since some of th e data did not include the start of green time (thus the headway for vehicle 1 could not be determined), this an alysis was based on a larger data set. The estimated coefficient values are shown in Tabl e 4-4. An R2 of 0.5923 was obtained for this model. 1 1 1 13 2 1 8 2 b Pct b Pct b Pct h hLT MT ST PC [4-5] Table 4-4. Results from STATISTICA for model 2 Coefficients Estimate Standard error t-value hPC 2.381300 0.021310 111.7433 b1 1.765421 0.099050 17.8236 b2 2.062019 0.118059 17.4660 b3 2.508828 0.077991 32.1682 Model 3 The third model (Equation 4-6) ha s the same characteristics as Model 1 but using only the data that contain the start-up time. The estimated coefficient values are shown in Table 4-5. An R2 of 0.3798 was obtained for this model. 1 1 1 13 2 1 b Pct b PctM b Pct h hLT ST ST SPC SAT [4-6] Table 4-5. Results from STATISTICA for model 3 Coefficients Estimate Standard error t-value hSPC 2.204078 0.040422 54.52684 b1 1.705201 0.189231 9.01121 b2 2.148228 0.217222 9.88953 b3 2.639080 0.174277 15.14304

PAGE 60

60 Model 4 The fourth model (Equation 4-7) ha s the same characteristics as Model 2 but using only the data that contain the start-up time. The estimated coefficient values are shown in Table 4-6. An R2 of 0.5309 was obtained for this model. 1 1 1 13 2 1 8 1 b PctLT b PctMT b PctST h hPC [4-7] Table 4-6. Results from STATISTICA for model 4 Coefficients Estimate Standard error t-value hPC 2.479479 0.024889 99.61976 b1 1.414669 0.102169 13.84634 b2 1.882030 0.117720 15.98733 b3 2.239560 0.093155 24.04133 These models resulted to have a low value for R2 and it was due to the variance inherent in the data. This could be due to insufficient field data. In addition, there were specified several other models that included addi tional variables such posted speed limit, area type, and number of lanes. These models were not f ound to be statistically significant. Start-up Reaction Time (SRT) The Start-up Reaction Time (SRT) is the time from start of green until the first vehicle begins to move. This measurement is sensitiv e to the timing precision of the video frame rate (0.0333 frames/sec). If vehicles in the first posit ion had their front axle on the stop bar, their headways were set equal to the SRT. Cycles th at had vehicles in the first position with their front axle beyond the stop bar were dropped out of the data set. These times were obtained from the videos and we re measured only from site 1. Site 1 has two through lanes, but these data were coll ected during a time when one through lane was closed. After careful observation and comparison between the queue discharge at this site for two through lanes and one through la ne, it was concluded that this particular lane closure set-up

PAGE 61

61 did not have any adverse impact on start-up react ion times or queue discharge rates. Right turning vehicles were not included in these measurements. A frequency graph of the field-measured SRT values measured is shown in Figure 4-1. 3 6 33 2626 11 12 8 4 2 6 2 1 3 11 2 11 0 1 00 1 0 5 10 15 20 25 30 35 1.251.752.252.753.253.754.254.755.255.756.256.75 SecondsFrequency Figure 4-1. Start-up r eaction time frequencies Based on the observations, it was decided th at a maximum of 3.5 seconds can be considered as a realistic maximum start-up reacti on time. SRTs greater than 3.5 seconds were typically due to an obvious hesitati on (related to distraction or sim ilar) of the drivers. Therefore, SRT values above 3.5 seconds were trimmed from the data set. Figure 4-2 shows a frequency distribution for only thos e drivers with a SRT of 3.5 seconds or less. 3 6 33 2626 11 12 8 4 2 0 5 10 15 20 25 30 35 1.251.51.7522.252.52.7533.253.5 SecondsFrequency Figure 4-2. Start-up r eaction time frequencies for trimmed data set

PAGE 62

62 Using this trimmed data set with 134 observa tions, a mean start-up reaction time of 2.04 seconds and a standard deviation of 0.47 were calculated. This mean value was somewhat consistent with previous studies reported in the literature. Kockelman [6] reported a mean SRT value of 1.79 sec and Li and Prevedouros [10] reported a mean SRT value of 1.76 and standard deviation of 0.61. They also concluded th at these values were normally distributed. For simulation purposes, it was assumed that th e first vehicle starte d at the stop bar and therefore the headway of the first vehicle in que ue was equal to the SRT regardless of vehicle type. From field data and other studies it was decided to the set SRT equal 2.0. Start-up Lost Time (SLT) The Start-up lost time is the difference between elapsed time for first four vehicles (1-4) to cross the stop bar and the time for the last four vehicles (5-8) in queue assuming passenger car only queue. The SLT can be calculated as it is shown in Equation 4-8. 44 1 SATh TT SLT [4-8] Where: TT1-4 = total time required for first four ve hicles in queue to cross the stop bar SATh = average saturation headway for a passenger cars only queue The field data were examined to obtain some sense of whether or not trucks are more or less likely to be in the first four positions of the queue. Table 4-7 summarizes these data. Table 4-7. Distributi on of trucks by site and position in queue Site indicator 1 2 3 4 5 6-1 6-2 6-3 6-4 Total Total cycles 50 11 24 17 22 52 66 18 143 403 Cycles with trucks 41 11 8 14 20 20 25 9 81 229 Cycles with trucks up-front* 6 1 3 8 11 9 10 7 40 95 % of trucks up-front* 15% 9% 38% 57% 55% 45% 40% 78% 49% 41% *up-front = in the first four positions

PAGE 63

63 Based on the field collected data it was not possible to draw a definiti ve conclusion about whether trucks are more or less likely to be at the front of the queue. Thus, for the purposes of this study, it was assumed that trucks were rando mly distributed throughout the entire queue, for each cycle. Additionally, it was observed that the Start-up Lost Time increased proportionally with an increase in overall truck percentage. Calibration of Simulation Model As mentioned before, for simulation to be effective, it is necessary to calibrate a simulation program against field data. An extensive calibra tion process was performed in order to identify the parameter values that resulted in the simulati on data providing the best match with the field data. The quantitative measure used in this process was the Mean Square Error (MSE). The MSE is commonly used to compare a predicted value with an observed value. In this case, it was used to compare the observed average headway fr om the field data with the average headway estimated from the simulation data. Equation 4-9 shows the mathematical form for MSE. N h h MSESimul Field2 [4-9] Where: MSE = Mean Square Error Fieldh = average headway from the field data Simulh = average headway from the simulation data N = total number of observations The calibration process consisted of compari ng valid field data cycles with simulation cycles with the same configuration. Valid cycles were those cycles with at least 8 vehicles in queue and that showed a normal behavior. Cycles with implicit hesitation of the drivers, lane changing, first vehicle beyond the stop bar, motorc ycles and right turns were considered as not valid or non-normal behavior.

PAGE 64

64 In this process there were a couple of parame ters that remained fixed. The simulation time-scan interval (T) is not a direct calibration paramete r and was not changed throughout the process. Also parameters as R and K were fixed at 0.7 and 1.25 respectively. These values were recommended by Cohen [11]. The vehicles lengths were fixed us ing values that were based on observations from this studys field data, and we re also consistent with values recommended by Long [22]. The rest of the parame ters were adjusted during the pro cess. A screen capture of the program with the final set of para meters is shown in Figure 4-3. Figure 4-3. User-adjustable vehicl e, driver, and model parameters In the process of calibration and trying to achieve a minimum MSE value, it was necessary to keep in mind that the final values of the parameters should stay within a reasonable range. For example, if a minimum MSE value is obtained by using a passenger car free flow speed of 30 ft/sec (~20 mi/h), it is not realistic to say that a passenger car will cruise at such a low speed. Process of calibration : An input file was created and read into the pr ogram to facilitate the process. This file contained each queue composition observed in the field. A first attempt included ten replication runs of each queue scenario. The variance within the runs was too high and the required sample size was on the order of 60 replicat ions, for that reason the number of replications was increased to 100. During the process of calibrati on the parameters that were primarily experimented with were the lead vehicles acceleration and the maximum acceleration, free-flow speed, minimum

PAGE 65

65 desired headway and inter-vehicle spacing per vehicle t ype. A total of 99 different parameter combinations were run in an effort to obtain the lo west MSE. In those iterations, the value of the MSE fluctuated from 0.097 to 0.415 for the aver age headway of vehicles 2 through 8 and from 0.142 to 0.761 for the average saturation headway. The final set of parameters had 0.118 and 0.142 for the average headway of vehicles 2 through 8 and the averag e saturation headway respectively. The final set of parameters is shown in Table 4-8. Table 4-8. Final choice of calibration parameter values Vehicle length Maximum acceleration FFS FFS std. dev. Headway Headway std. dev. Stop gap Stop gap std. dev. (ft) (ft/s2) (ft/s) (ft/s) (sec) (sec) (ft) (ft) PC 15 10 72.50 3.75 1.50 0.25 10 2.0 ST 30 5 67.50 3.75 2.50 0.25 14 2.0 MT 45 4 62.50 3.75 3.00 0.25 16 2.5 LT 65 3 57.50 3.75 3.50 0.25 20 2.5 Execution of Experimental Design Once the experimental design was complete, an input file was gene rated with the 7459 queue combinations. The simulation program was run with this file pe rforming 100 replications of each queue combination. Based on some testi ng, it was decided to perform 100 replications to account for the considerable variance inherent in this process. Since there was considerable variance in the field data, the ca libration process of the simulation still allows for a large amount of variance in each individual simulation run. The choice of 100 replications was somewhat conservative, but not extremely conservative. An average of each of the 100 replications was calculated and used in the analysis database. This new data set was inputted in the statistical software package STATISTICA. Subsequently, variables were created as needed to run the models. These variables include, but are not limited to, indicator variables for each vehi cle type per position, indicator variables for vehicle pairs per position, frequencies and percentages of each one of theses scenarios.

PAGE 66

66 Additional experiments were run to verify certain numerical re sults, such as the following. 10,000 passenger car-only queues to veri fy the base saturation headway 400 passenger car-only queues, changing the vehicl e in position 4 to estimate the effect of different truck types on the headwa y of the vehicle in position 5. 256 cases with 100 replications of trucks in positions 1-4 and PCs in positions 5-8 to isolate the effect of trucks on the st art-up lost time section of the queue. A data set with fabricated fixed values for each vehicle pair, regardless of their queue position, to verify model formulations.

PAGE 67

67 CHAPTER 5 ANALYSIS OF SIMULATION DATA This chapter describes how the data genera ted from the simulation experimental design process were used to obtain the new PCE factors. For the purposes of this research, the queue was analyzed in sections that were consiste nt with the guidance given in the HCM 2000 with regard to the components of start-up lost time a nd saturation headway. Th at is, the first four headways are included in the start-up lost time, and the headways of the remaining vehicles represent the saturation headway. Thus, the queue was subdivided into two sections for purposes of the analysis. Figure 5-1 illustrates these two queue sections. 8 7 6 5 4 3 2 1 Start-Up Lost Time Section Saturated Section of Queue Figure 5-1. Queue sections for analysis In reality, the lost time does not necessarily e nd with the headway of the fourth vehicle, but for purposes of this research it was decided to be consistent with the HCM framework. Furthermore, the field data collected in this study were not conclusive with regard to this issue. As previously mentioned, the analyses we re done assuming trucks were distributed randomly within each queue. That is, for any gi ven cycle, a given number of trucks may appear at any location throughout the queue. In some qu eues, a given number of trucks may be at the front of the queue, in some queues this same num ber may be in the middle of the queue, and so on. But over the length of the analysis period, fo r a fixed percentage of trucks in the traffic stream, it is assumed that any position within the queue is as likely to have a truck in that position as any other position in the queue.

PAGE 68

68 The first approach to find the PCE values was to analyze the saturated part of the queue, as it is recommended in the HCM. A simple model us ing only the percentages of each type of truck in positions 5 through 8 was specified. The resu lts of this model did not reflect exactly the behavior of the queue. This was expected sin ce this model does not explicitly account for the headways being a function of the specific leader -follower combination. Subsequently, another model was specified using the frequency of each ve hicle pair in the second half of the queue. This approach assumed that each vehicle pair had the same headway regardless of their position in the queue. The estimates from this model di d not replicate the performance of the queue as closely as hoped. To confirm that the model was specified correctly, a different program was developed to generate fixed headway values that were independent of queue position. When the same model was run on this data set, it provided a better fit, but still not exact. Thus the next step was to incorporate indicator variables fo r each one of the 16 different vehicle pair combinations, for positions 5 through 8. This mo del demonstrated that each vehicle pair has a different headway value dependi ng of their position in queue. It was also noted that the type of vehicl e in position 4 had a si gnificant impact on the calculation of saturation headway, as expected. Furthermore, it was also observed that trucks present in the first few positions of the queue also influence the saturation headway. Figure 5-2 shows a diagram with three different scenarios reflecting these impacts. Figure 5-2. Impact of trucks in the first part of the queue to the saturation headway The first case (on top) has only passenger cars the second case (middle) has a large truck in position 4, and the third case (bottom) has la rge trucks in positions 3 and 4. Although the

PAGE 69

69 second part of the queue was the same for all the cases (passenger cars in positions 5-8), the saturation headways were significantly different. The saturation headways for these three scenarios were 2.033, 2.795 and 2.641 seconds re spectively. The only difference between the first and second cases is the presen ce of a large truck in the four th position. The size of this vehicle impacts directly the headway of the vehicle in position 5. In addition, large trucks have a lower acceleration rate when compared to passe nger cars, which affect the acceleration of the vehicles behind. In this case, vehicles in pos itions 5 through 8 were directly impacted by the truck. These vehicles were constrained to the acceleration of the large truck since lane changing was not allowed. This comparison confirms th at the saturation headway is affected by the vehicle in position 4. Moreover, it was observed how the impact on saturation headway due to the number of trucks in the first few queue positions might be overestimated. The impact on saturation headway is just as much a function of the position of the trucks in the queue as it is the total number of trucks in the first f our positions of the queue. These data showed that the queue with two large trucks had a smaller saturation headwa y than the one with only one truck. This confirms that not only the vehicle type in positio n 4 impacts the saturation headway, but also the presence of other trucks preceding the truck in po sition 4 provide an additional impact. In this example, the difference results from the length of the trucks rather than their acceleration capabilities. Vehicles in positions 5 through 8 we re constrained to the acceleration rate of the large trucks, but in the third case the vehicles were farther back in the queue. The distance to the stop bar of each vehicle was larger and with the same acceleration the vehicles reached higher speeds prior to crossing the stop bar. The spacing was very similar between vehicles in the two cases, so with a higher speed and similar spacing, the result is lower headways. This result is

PAGE 70

70 consistent with some studies that have claimed th at the headways between vehicles at the back of a long queue will be lower than those near the fr ont because the vehicles near the back have more time accelerate to a higher speed. However, the Li and Prevedouros [10] study found that this is not always the case because headway expa nsion may occur toward the end of long queues In light of these findings, it was desired to test a model that included indicator variables for each one of the 16 different vehicle-pair combinations for each queue position. However, this results in a model with an extremely large number of parameters, which was beyond the capabilities of the statistical soft ware package. Therefore, it was decided to treat each part of the queue separately. A new data set was created with only passenger cars in the second part of the queue (positions 5-8). This data set included 256 (44) different queue combinations, as it was only the first four positions of the queue that varied by vehicle type. This new data set was used to estimate the impact of trucks to the SLT. The more accurate method to analyze the impact of trucks to the SLT is to use indicator variables for the 16 different vehicle-pair combinations in each one of the first four positions. But again, this approach was not practical. Th erefore, since the simula tion program provides the same headway value for the first vehicle regardless of what type it is, it was decided to run a narrower model with only indicator variables (for the 16 combinati ons) for the first two vehicles. This model isolates the impact of the first vehicl e in queue. The headway of the first vehicle was the same for each vehicle type, but the effect to the trailing vehicle due to the length and acceleration of the first vehicle va ried. After the results from this model were obtained, this model was included in another model that included indicator variables of vehicle types for in positions 2 through 4. This model is shown in Equation 5-1. 4 2 4 3 2 4 3 2 1 4 1k k k k PCLT b MT b ST b TT TT TT [5-1]

PAGE 71

71 Where: TT1-4 = clearance time for the first four vehicles in queue TT1-2 = clearance time for the fi rst two vehicles in queue TTPC3-4 = clearance time for passenger cars in positions 3 and 4 bik = additional time for vehicle type i in position k to clear ST = indicator variable for vehicle type small truck (1-yes, 0-no) MT = indicator variable for vehicle type medium truck (1-yes, 0-no) LT = indicator variable for vehicle type large truck (1-yes, 0-no) k = indicator for queue position i = ST for small truck, MT for me dium truck, LT for large truck And the clearance time for the first two vehicles was defined as is s hown in Equation 5-2. 4 1 2 10 2j i ij ij PC PCI b h TT [5-2] Where: TT1-2 = clearance time for the fi rst two vehicles in queue hPC-PC = headway of a passenger car in position 2 following a passenger car bij = additional time for vehicle i following vehicle type j to clear Iij = indicator for vehicle type i following vehicle type j i = indicator for vehicle type j = indicator for vehicle type The final form of the equation to estimate the to tal time to clear the first four vehicles in queue (Equation 5-1) is shown in Equation 5-3. 44 43 42 34 33 32 244 243 242 241 234 233 232 231 224 223 222 221 214 213 212 4 1b 1.72 b 1.21 b 0.83 b 4.03 b 2.49 b 1.47 b 47 9 b 36 7 b 96 5 b 22 4 b 80 7 b 65 5 b 19 4 b 47 2 b 65 6 b 53 4 b 09 3 b 38 1 b 5.04 b 3.00 b 1.62 10.59 TT [5-3] Where the total time to clear th e first four vehicles, if are passenger cars, is 10.59 seconds and the bkij and bij follow the format described earlier in this chapter. From Equation 4-8, the SLT can be calculated as the difference of this time to clear the first four vehicles and the equiva lent of four times the saturati on headway of passenger cars. The SLT calculated for passenger cars only queue is equal to 2.47 seconds, which is higher than the

PAGE 72

72 2.0 seconds recommended in the HCM 2000 [1], but somewhat consistent with the findings of the field data. It is also important to account for the impact of trucks to the S LT since the presence of these vehicles increased significantly the SLT. In example, if the composition of the queue is as it is shown in Figure 5-3. Figure 5-3. SLT example where a large truck is in the second position, a medium truck in the fourth position, and the rest are passenger cars, the SLT can be estimated as indicated in Equation 5-4. 03 2 4 b 1.72 b 22 4 10.5944 241 SLT [5-4] Then, the SLT for this case of mixed traffic is 8.41 seconds. This value is considerably higher when compared to the 2.47 seconds for passenger cars only queue and it needs to be treated separately from the saturation headway. Since it is not realistic to e xpect the practitioner to collect every vehicle type in queue and their exact position, an equation was developed to estimate the SLT based on the percentage of trucks in the queue. As discusse d earlier, it is assumed that th e trucks are distributed randomly throughout the queue. Therefore, over many cycles with trucks distribu ted randomly within the queue for each cycle, it can be assumed that on average, trucks will be distributed evenly throughout the queue. The SLT equation is given by, LT MT STPct Pct Pct SLT 0 15 0 9 0 5 5 2 [5-5] Where: Pcti = percentage of truck type i

PAGE 73

73 After analyzing the two different sections of queue, a compound model was developed. This model was used to obtain an estimate of th e total time elapsed for clearing a queue length of eight vehicles, using the experiment al design data set. A general form of this model is shown in Equation 5-6. 18144() P CPCijijTTTThbF [5-6] Where: TT1-8 = clearance time for the firs t eight vehicles in queue TT1-4 = clearance time for the firs t four vehicles in queue hPC-PC = headway of passenger car following a passenger car bij = additional headway for vehicle pair (vehicle i following vehicle j) Fij = frequency of vehicle pair (vehicle i following vehicle j) in positions 5 through 8 The model estimation output is shown in Table 5-1 where b0 is hPC-PC and the remaining betas follow the bij format described before. Table 5-1. Model estimation results for additio nal headways of 16 vehicle pair combinations Coefficients Estimate Std. error t-value hPC-PC 2.028586 0.004183 484.9803 b12 0.590274 0.012212 48.3346 b13 1.046166 0.012212 85.6654 b14 1.852489 0.012212 151.6913 b21 1.024162 0.012212 83.8636 b22 1.531207 0.016902 90.5956 b23 2.035979 0.020033 101.6312 b24 2.993811 0.020033 149.4440 b31 1.407661 0.012212 115.2665 b32 1.920917 0.020033 95.8877 b33 2.393117 0.016902 141.5915 b34 3.377253 0.020033 168.5845 b41 1.823759 0.012212 149.3388 b42 2.426886 0.020033 121.1444 b43 2.835042 0.020033 141.5186 b44 3.573751 0.016902 211.4451

PAGE 74

74 These estimates were substituted in Equation 5-6 and the final form of the model is defined in Equation 5-7. 44 43 42 41 34 33 32 31 24 23 22 21 14 13 12 4 1 8 1F 3.574 F 2.835 F 2.427 F 1.824 F 3.377 F 2.393 F 1.921 F 1.407 F 2.994 F 2.036 F 1.531 F 1.024 F 1.852 F 1.046 590 0 029 2 4 F TT TT [5-7] Figure 5-4 shows a plot of this equati on against the observed simulation data. 18 20 22 24 26 28 30 32 34 36 38 18202224262830323436 ObservedPredicted Figure 5-4. Total time for vehicles 1-8 using vehicle pairs in positions 5-8 This model provides the headway for a pa ssenger car following a passenger car (hPC-PC) in positions 5 through 8 and also the additional time headway taken by any other vehicle pair in these positions. Therefore, it was possible to ca lculate the headway for each one of these pairs. Table 5-2 shows these headways. In this table the PCE value for each vehicle pair is also shown. These PCE values were calculated as a relative headway to the hPC-PC per the definition in the HCM 2000.

PAGE 75

75 Table 5-2. Headways and PCE values for 16 vehicle pair combinations Vehicle pair trailing leading hPC-PC (seconds) Additional headway* (seconds) Headway (seconds) PCE values of each 16 vehicle-pair PC PC 2.029 2.029 1.000 PC ST 0.590 2.619 1.291 PC MT 1.046 3.075 1.516 PC LT 1.852 3.881 1.913 ST PC 1.024 3.053 1.505 ST ST 1.531 3.560 1.755 ST MT 2.036 4.065 2.004 ST LT 2.994 5.022 2.476 MT PC 1.408 3.436 1.694 MT ST 1.921 3.950 1.947 MT MT 2.393 4.422 2.180 MT LT 3.377 5.406 2.665 LT PC 1.824 3.852 1.899 LT ST 2.427 4.455 2.196 LT MT 2.835 4.864 2.398 LT LT 3.574 5.602 2.762 h = (Vehi Vehj)-(PC PC) Although these PCE values reflected more accura tely the behavior of the queue, it is not practical to have a PCE value for each possible vehicle-pair combina tion. In an effort to simplify this model, a new set of PCE factors was derived for each one of the truck vehicle types. This new set takes into consideration their headway and the time each vehicle type adds to the following vehicles headway. This concept can be described as the time consumed by any vehicle type in queue. Figure 55 illustrates this concept with an example of a Large Truck in queue. This example compares the headway of a PC following a LT with the headway of a PC following a PC. Figure 5-5. Time consumed by a large truck

PAGE 76

76 As shown, a LT adds an additional headway ( h) to the PC. In this case the additional time can be defined as it is shown in Equation 5-8. PC PC LT PC LTh h h [5-8] From the values in Table 5-2, 1.852 seconds is added to the headway of the passenger car when following a large truck. This additional time should be considered part of the LT headway rather than the PC since this additional time it is not present in a PC only queue. Also shown is the time consumed by a LT in queue (dark shade) which was defined as the time headway of a LT plus the additional headway this vehicl e type adds to the following vehicle. The h of each vehicle type is not only added to the PC. These impacts are present regardless of what type of vehicle is following. Table 5-3 includes a summary of each vehicle types headway when following a PC and the h they add to any other vehicle type. Table 5-3. Vehicle t ypes headway and their h Leading Trailing Headway when following PC h ST h MT h LT PC 2.029 0.590 1.046 1.853 ST 3.053 0.507 1.012 1.970 MT 3.436 0.513 0.986 1.970 LT 3.852 0.603 1.011 1.750 Average h 0.553 1.014 1.885 In order to develop a model that does not ta ke into consideration the position of each vehicle in queue, it was n ecessary to use the average h per vehicle type. By using the average of the h, the proportion of these vehicl e pairs in queue are not being taken into consideration. To develop PCE values based on just three truc k categories, rather than all 16 possible leadfollow vehicle pairings, the resu lting PCE values will obviously be generalized. That is, with this approach, information about specific posi tion in queue for each truck is not utilized. However, the reason, again, for this approach is that it is expected that practitioners will not collect data at this level of detail. With such a generalized approach, this loss of information

PAGE 77

77 requires some level of approximation. The implic it assumption is that each vehicle pair occurs with the same frequency in the traffic stream. While this may rarely be the case, the objective was to arrive at generalized PCE values that yielded an fHV value that still tracked reasonably well with the fHV value that results from using all 16 PC E values, for a varying range of overall truck percentage, as well as vary ing relative truck type percentages. This more general equation using the time consumed (H) by any vehicle i in queue is defined in E quation 5-9. Where vehicle j is following vehicle i. i PC i ih h H [5-9] Where: 44 1 j i PC j i j ih h h [5-10] And 1 = PC, 2 = ST, 3 = MT and 4 = LT Shown in Table 5-3 are the values for the time consumed in queue by each vehicle type. These values can be used to arri ve at new PCE factors since this time consumed by each vehicle type can be considered as their time headway. Table 5-4 shows the PCE factors that were calculated as the relative time consumed per each vehicle type in queue when compared to a PC. Table 5-4. Time consumed and PCE values for each vehicle type Vehicle type hi PC (seconds) Avg. hi (seconds) Hi (seconds) PCEi PC 2.029 0.000 2.029 1.000 ST 3.053 0.553 3.606 1.778 MT 3.436 1.014 4.450 2.194 LT 3.852 1.885 5.738 2.828 To assess the accuracy of these values, a model of total time needed to clear a queue with eight vehicles with only three t ypes of trucks, instead of all vehicle-pair combinations, was tested. The model was specified in STATISTICA and the general form of this new model is shown in Equation 5-11. LT LT MT MT ST ST PCF b F b F b h TT TT 8 5 4 1 8 14 [5-11]

PAGE 78

78 Where: TT1-8 = time needed to clear the fi rst eight vehicles in queue TT1-4 = time needed to clear the fi rst four vehicles in queue hPC5-8 = headway of passenger cars followi ng passenger cars in positions 5 through 8 bi = additional headway for vehicle type i Fi = frequency of truck type i in positions 5 through 8 i = ST for small trucks, MT for medium trucks, and LT for large trucks The model estimation results are shown in Table 5-5, where b0 represents the time of only PCs in positions 5 through 8 and the remaining betas follow the bi format described before. Table 5-5. Results from STATISTI CA for model with vehicle types Coefficients Estimate Std. error t-value hPC5-8 2.270076 0.029382 309.0408 bST 1.228277 0.018823 65.2534 bMT 1.952747 0.018823 103.7416 bLT 2.985894 0.018823 158.6285 Figure 5-6 shows the plot of this equation against the observed data. 18 20 22 24 26 28 30 32 34 36 38 18202224262830323436 ObservedPredicted Figure 5-6. Total time for vehicles 1-8 using vehicle types in positions 5-8

PAGE 79

79 This figure illustrates the increase in varian ce when compared to the plot based on all 16 vehicle pairs, as expected. Usi ng just the three general categories of truck type instead of all 16 vehicle pairs obviously results in less precision of the saturation fl ow rate estimate. Table 5-6 shows the resulting headways and PCE values calc ulated from the results shown in Table 5-5. Table 5-6. PCE factors for three vehicle types Vehicle type headway (seconds) PCE PC 2.270 1.000 ST 3.498 1.541 MT 4.223 1.860 LT 5.256 2.315 Note that these PCE values are considerably lower than the ones calculated from the other model shown in Table 5-4. These values demons trate that the impact of the passenger cars is being overestimated and the impact of the truc ks is being underestimated. As previously discussed, a truck in position four has a significant impact to the saturation headway, and this approach did not take th at into consideration. Another alternative was to us e the alternate form of the fHV equation from the HCM. The following derivation was made with the purpose of relating the current analysis to the HCMs approach. Starting with the model specified to es timate the total time needed to clear the first eight vehicles in queue Equation 5-11, and if 8 5 4 1 8 1 TT TT TT then, LT LT MT MT ST ST PCF b F b F b h TT 8 5 8 54 [5-12] Again, the HCM defines the PCE as a relative headway. Since hPC5-8 is the headway of a passenger car and bi is the additional head way of vehicle type i, then 8 5 8 5 8 5 8 5 PC PC i i PC PC i ih h PCE b h h b PCE [5-13]

PAGE 80

80 Then, LT PC LT PC MT PC MT PC ST PC ST PC PCF h PCE h F h PCE h F h PCE h h TT8 5 8 5 8 5 8 5 8 5 8 5 8 5 8 54 [5-14] If, 48 5 TT hSAT [5-15] and 4i iF Pct [5-16] then simplifying the equation and dividing by 4 to obtain one average headway value yields, 1 1 1 18 5 LT LT MT MT ST ST PC SATPCE Pct PCE Pct PCE Pct h h [5-17] This model was used to generate the PCE values. The results were similar to those obtained when using vehicle type s in the total time formulation (Equation 5-11). This model was also run with a fixed value for hPC-PC and the results were similar to the ones obtained when using vehicle pairs (Equation 5-7). From these results it was concl uded that the impact of trucks in the first few positions of the queue was not ac counted for. In addition, the data set was expanded adding passenger car only queues to redu ce the overall truck percentage from 46% to 10%. Then the same model was run (Equation 5-11) and the results were the same as fixing the hPC-PC. Increasing the amount of passenger car only queues in the data set had a significant change to the estimation of the saturation headway. The alternate form of the equation for fHV can also be derived from the model based on all 16 vehicle pairs. Equation 5-18 gives the final form of this equation. 4 1 8 51 1j i ij ij PC PCPCE Pct h h [5-18]

PAGE 81

81 Model estimation results are shown in Table 5-7, where b0 estimates the hPC-PC and the remaining betas estimate the PCE for vehicle i following j. Table 5-7. Model estimation results for fHV equation with 16 vehicle pairs Coefficients Estimate Std. error t-value hPC-PC 2.099136 0.002592 809.6979 b12 1.255759 0.003755 334.4511 b13 1.474404 0.003898 378.2679 b14 1.860251 0.004181 444.9291 b21 1.431489 0.003869 370.0254 b22 1.672034 0.005277 316.8362 b23 1.910199 0.006270 304.6783 b24 2.364296 0.006512 363.0607 b31 1.614128 0.003996 403.9260 b32 1.855496 0.006243 297.2072 b33 2.079154 0.005505 377.6886 b34 2.545657 0.006620 384.5518 b41 1.813318 0.004145 437.5067 b42 2.097403 0.006365 329.5371 b43 2.290151 0.006470 353.9746 b44 2.638246 0.005874 449.1694 Although this form is consistent with the HCM, to base the new PCE values it was decided to use the model (Equation 5-7) that estimates the total time for the first eight vehicles in queue to clear with the 16 different vehicle pairs. Accounting for the total time provides a better understanding of what occurred in the entire queue and it isolated the impact of trucks to the SLT from the saturation headway. However, the main objective of this research was to develop new PCEs for three truck cate gories that result in an fHV value that match reasonably well with the values based on all 16 PCEs. It is not possible to have only one set of PCE values (for just the three truck categories) and have them result in the same fHV value as for the 16 PCE values across a range of truck percentages. In order to obtain the new PCE values for th e three trucks types from the sixteen vehicle pairs headway, two different methods were implem ented. The first was to use the relative time

PAGE 82

82 consumed in queue (H) by each vehicle type. This method considered only the headway of each vehicle i when following a passenger car plus the average additional head way of any vehicle j trailing vehicle i. The results of this method are shown earlier in this chapter in Table 5-4. The second was implemented adjusting the addi tional time consumed by each vehicle in queue. This adjustment comes from the obser vation that the time consumed by a vehicle i it is not exactly the same across different trailing vehicl es. A general for of this method is shown in Equation 5-19. 34 2 j i j j i i PC i i adjh h h h H [5-19] Where; Hadj-i = adjusted time consumed by vehicle i in queue hi,PC = headway of vehicle i when following a PC ih = average additional headway by vehicle i j ih, = additional headway of vehicle i when following vehicle j Using the same approach of relative time consumed, PCE values of 1.75, 2.16, and 2.80 were obtained for small trucks, medium trucks, and large tr ucks respectively. These values were extremely similar, and to be consistent with these results, it was decided to use values of 1.8, 2.2, and 2.8 for small, medium and large trucks respectively. Tests done with these values across a wide range of overall truck percentage, as well as relative truck type percentages, showed that the resulting fHV values tracked reasonably well with those yielded by applying all 16 possible vehicle l ead-follow PCE values. For situations where the relative truck type percentages might be extremely skewed (such as an exit from a distribution warehouse that consists entirely of large trucks) the specific PCE value from Table 5-2 can be used directly (in this case, PCELT-LT).

PAGE 83

83 Furthermore, when relative truck type di stributions are not available, a general approximation based on a relatively balanced distribution of small, medium, and large trucks in the traffic stream can be assumed. This approx imation consisted of the average of the PCEs for these three categories. From Table 5-4 PCE values of 1.8, 2.2 and 2.8 can be obtained for small, medium and large trucks respectively. The aver age of these values yields to 2.267. Thus, a single truck PCE value of 2.3 can be applied. Again, this rough estimate considered a general assumption that small trucks, medium trucks and la rge trucks are equally distributed in the traffic stream.

PAGE 84

84 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS The signal analysis methodology (Chapter 16 ) of the Highway Capacity Manual (HCM) currently recommends a single truc k PCE value of 2.0. There is not much research available to support this value, and some tran sportation professionals have que stioned the validity of this value, feeling that it generally unde restimates the impact of trucks on saturation flow. This issue was the focus of this study; thus, the objective was to either validate this specific PCE value from the HCM, or to recommend a more appropr iate value, or combination of values. The results of this study are based primarily on simulation data. While not as ideal as having the results based strictly on field data, for this type of study that was focused on the effects of large trucks, there we re some additional constraints beyond the typical saturation flow rate study that does not consider the effect of trucks. The challenge of finding sites with long queues (eight or more vehicles) that also had a large percentage of trucks, the large variance of the effect of trucks, and the sample size requir ed due to this variance made it prohibitive to collect sufficient field data in an efficient manne r within the scope and re sources of this project. Nonetheless, a considerable amount of field da ta were still collected which provided for a reasonable data set to use for simulation program calibration. It was felt that the simulation program was calibrated fairly well to the field da ta, and replicated the ge neral trends observed in the field data quite well. Thus, it is felt that the results provided in this study, despite the heavy reliance on simulation data, are re asonably valid and reliable. The results presented in this study were base d on using the same definition for saturation headway as provided in the HCM; that is, the av erage headway of vehicles in position 5 through 8 of the queue. Some studies [10] have indicat ed that the saturation flow rate will increase (saturation headway will decrea se) with longer queues; althou gh the body of evidence at this

PAGE 85

85 point is inconclusive. Li and Prevedouros [10] also found that in some cases the saturation headway may increase, and in others it might d ecrease. They indicated that in the former, headways at the end of a long queue compress, a nd in the latter headways elongate, and this is a function of drivers performance. In this study, it was found that trucks in the first few positions of the queue impact not only the start-up lost time, but also the saturation headway of short queues. This result likely reflects the fact that start-up lost time extends beyond the fourth vehicle in queue, as it is expected that satura tion headway would eventually reach a stable value if the queue is sufficiently long enough. This result could not be confirmed directly, due to computational limitations of the simulation program. As indicated before, for purposes of this rese arch it was assumed that the saturation section of the queue started after the four th vehicle. Therefore it was deci ded to exclude this increase in the saturation headway because it is most likely a result of the a dditional lost time caused by the trucks being absorbed into the saturation flow ra te estimate for short queues. From a practical standpoint, the accuracy of capacity estimates is not critical for situations when queues at signalized intersections ar e consistently short. As it was demonstrated, headways are a func tion of both the leading and following vehicle in the queue, not just the trailing vehicle. This study categorized vehicles into four different types, which led to a total of 16 (42) possible leading-trailing vehi cle pairs. New PCE values were developed for these pairs; however, it is not realistic to expect a practitioner to collect such detailed data on vehicle pair fr equencies to apply these PCE values. Therefore, in order to provide more practical results, PCE values were developed based on three general categories of truck typesmall, medium, and large. The method by which this was accomplished was to consider the time each vehicle type consumed during the queue clearance process. This time

PAGE 86

86 consumed was defined as the headway of the vehicl e plus an additional time it adds to the trailing vehicle, and was based on the values ob tained from the model that considered all 16 possible lead-trail vehicle pairs. The final reco mmended values for these PCEs are listed below. 1.8 for small trucks 2.2 for medium trucks 2.8 for large trucks In this study small trucks in clude those trucks that have only two axles and between four and six tires. It also includes passenger cars wi th a trailer and garbage tr ucks, regardless of their number of axles. Medium trucks include those with three axles and usually range in length from 40 to 55 ft. Passenger cars with a trailer using the fifth-wheel, recreational vehicles (RVs) and small trucks with a trailer are also included as medium trucks. Large trucks include those with four or more axles, RVs with trailers and buses. For situations in which only an estimate of the overall percentage of trucks in the traffic stream is available, then the single truck PCE value of 2.3 can be applied. Again, this value is a very general approximation, based a relatively ba lanced distribution of small, medium, and large trucks in the traffic stream. Use of this single general truck PCE value is reasonable for planning applications, where the level of precision may no t be as important, or where it may not be possible to know the percentage s of different truck classifica tions in the traffic stream. For situations where the truck percentages, by category, are very unbala nced, the values of Table 5-2 can be used instead of the generalized ones. For example, if an approach to a signalized intersection serves a warehouse distribution center, where almost all the trucks are large, then the PCE value for LT-LT could be applied in the fHV equation. The increase in SLT due to trucks in the first four positions of the queue can be estimated using the percentage of trucks in the traffic stream. This estimate was based on the assumption

PAGE 87

87 that the trucks are distributed evenly throughout the queue. The re sults of this study showed a SLT for passenger car-only queues of approximately 2.5 seconds, which is larger than the HCM recommended value of 2.0 seconds. This value in creases accordingly with an increase in truck percentage, reaching a value of approximately 17. 5 seconds for 100% large trucks in the traffic stream. Another finding of this study relates to the HCM recommended base saturation flow rate value of 1900 pc/h/ln. For the field data obtaine d in this study that included queues of only passenger cars, this saturation flow was rarely ever reached. As th e results indicated, the average saturation headway for passenger car only que ues was approximately 2.03 seconds, which corresponds to a saturation flow rate of 1773 pc/hr. While it is certainly debatable as to whether any of the field sites had what would be considered truly ideal conditions, per the general HCM definitions, this was generally the case. On ly under the most optimal driver behavior and vehicle composition (e.g., only small passenger ca rs) conditions were average saturation headways below 2.0 observed. Although this fiel d data set was not extremely large, the HCM recommended value for base satura tion flow rate appears to be quite optimistic. This would seem to be particularly the case where the perc entages of SUVs, mini-vans, and pick-up trucks currently make up a significant percentage of passenger cars." Research by Kockelman and Shabih [6] found average headway values for thes e vehicle types to be greater than those for sedan-type passenger vehicles. From detailed observations of the field data, it was found that a very high percentage of discharging queues have one or more drivers th at hesitate/lag during th eir start-up process. While difficult to conclude exactly why this ha ppens from the video re cordings, it appears to frequently be a result of general driver inattentiv eness or distraction. That is, the phenomenon of

PAGE 88

88 drivers falling asleep at the wheel while waiting to startup from a stop at a signalized intersection approach appears to be quite comm onone or two drivers (within the first eight positions of the queue) almost every cycle. Th ese hesitating/lagging drivers have a significant impact on the capacity of the inters ection. Queues that contained obvious instances of this driver hesitation were not included in th e calculated value of 2.03; thus, this value still represents more ideal driver behavior. Unfortuna tely, this ideal driver behavior, at least for an entire discharging queue, seems to be quite rare. So based on just this phenomenon, the HCM recommended base saturation flow rate of 1900 pc/hg/ln would appear to be essentially unattainable over any reas onable length of time. Recommendation for further studies Ideally, a very large field study should be conducted, including not only all the variables in Bonnesons study [13], but also a comprehensive consideration of trucks and other vehicle types affecting the performance of the intersections. For example, heavily loaded trucks obviously have poorer acceleration capabilitie s than an unloaded truck. Ho wever, it must be recognized that trying to incorporate truc k weight information into the methodology may not be practical, due to the difficultly associated with obtaining th is type of measurement. With respect to determining the distribution of trucks within th e queue (i.e., by queue pos ition), a larger field study might also provide for the ability to reach a statistically valid conclusion on this issue. A study including longer queues should be done in order to better iden tify the length of queue over which the Start-up Lost Time applies when trucks are present at the front of the queue. From the results of this research, truc ks in the first four queue positions impact the discharge rate of vehicles in positions five th rough eight, and to better estimate the saturation flow rate it was necessary to exclude this im pact. Studying longer queues in the field will also provide more insight into the issu e of queue compression or elongation.

PAGE 89

89 It is also recommended to st udy the frequency of inattentive drivers within the queue. It was observed how these situations frequently resulted in large ga ps between vehicles, therefore resulting in an effective reduc tion of intersection capacity. This driver inattentiveness phenomenon could be could be included as a fa ctor in the adjusted saturation flow rate calculation, somewhat analogous to a local adjustment factor.

PAGE 90

90 APPENDIX A DATA COLLECTION EQUIPMENT SETUP Figure A-1. Data collection equipment setup for method 1

PAGE 91

91 Figure A-2. Signal controller cabinet with da ta collection equipment installed for method 1

PAGE 92

92 APPENDIX B PICTURES OF VEHICLE TYPES BY CATEGORY A B C D Figure B-1. Small trucks. A) Panel truck. B) Garbage truck. C) Two-Axle Single-unit dump truck. D) Small delivery truck. E) Passenger cars with trailers

PAGE 93

93 E Figure B-1. Continued

PAGE 94

94 A B C D E Figure B-2. Medium trucks. A) Three-Axle Single-unit dump truc k. B) Concrete Mixer. C) Passenger car with trailer using fifth wheel. D) Delivery truck. E) Single-unit cargo truck.

PAGE 95

95 A B Figure B-3. Large trucks. A) Tractor plus trailer. B) Tract or plus flatbed. C) Buses.

PAGE 96

96 C Figure B-3. Continued

PAGE 97

97 APPENDIX C SIMULATION PROGRAM SCREENSHOTS A B Figure C-1. Simulation screenshot A) Before signal turns green. B) Once the queue starts to discharge

PAGE 98

98 APPENDIX D QUEUE COMPOSITION Table D-1. Vehicle type per position in queue Freq Position in queue 1st 2nd 3rd 4th 5th 6th 7th 8th 1 174 1 1 1 1 1 1 1 1 2 2 1 1 1 1 1 1 1 2 3 3 1 1 1 1 1 1 1 3 4 9 1 1 1 1 1 1 1 4 5 9 1 1 1 1 1 1 2 1 6 6 1 1 1 1 1 1 3 1 7 6 1 1 1 1 1 1 4 1 8 5 1 1 1 1 1 2 1 1 9 5 1 1 1 1 1 3 1 1 10 11 1 1 1 1 1 4 1 1 11 5 1 1 1 1 2 1 1 1 12 6 1 1 1 1 4 1 1 1 13 6 1 1 1 2 1 1 1 1 14 1 1 1 1 4 1 1 1 1 15 3 1 1 2 1 1 1 1 1 16 3 1 1 3 1 1 1 1 1 17 10 1 1 4 1 1 1 1 1 18 7 1 2 1 1 1 1 1 1 19 3 1 3 1 1 1 1 1 1 20 7 1 4 1 1 1 1 1 1 21 6 2 1 1 1 1 1 1 1 22 6 3 1 1 1 1 1 1 1 23 7 4 1 1 1 1 1 1 1 24 1 1 1 1 1 1 1 2 2 25 1 1 1 1 1 1 1 3 2 26 1 1 1 1 1 1 1 3 4 27 6 1 1 1 1 1 1 4 4 28 1 1 1 1 1 1 2 3 1 29 1 1 1 1 1 1 2 4 1 30 1 1 1 1 1 1 2 1 4 31 1 1 1 1 1 1 3 2 1 32 1 1 1 1 1 1 4 2 1 33 1 1 1 1 1 2 1 2 1 34 2 1 1 1 1 2 1 1 4 35 2 1 1 1 1 2 2 1 1 36 1 1 1 1 1 3 1 2 1 37 1 1 1 1 1 3 1 4 1 38 1 1 1 1 1 3 3 1 1 39 1 1 1 1 1 4 1 2 1 40 1 1 1 1 1 4 1 1 2 41 1 1 1 1 2 1 1 2 1 42 1 1 1 1 2 1 1 4 1 43 1 1 1 1 2 1 1 1 4 44 1 1 1 1 2 3 1 1 1 45 1 1 1 1 4 1 1 2 1 46 1 1 1 1 4 1 1 3 1 47 1 1 1 1 4 1 1 4 1 48 2 1 1 1 4 1 1 1 3 49 1 1 1 1 4 1 4 1 1 50 1 1 1 2 1 1 1 3 1 51 2 1 1 2 4 1 1 1 1 52 1 1 1 4 1 1 1 1 2 53 1 1 1 4 1 1 3 1 1

PAGE 99

99 Table D-1. Continued Case Freq Position in queue 1st 2nd 3rd 4th 5th 6th 7th 8th 54 1 1 1 4 1 4 1 1 1 55 3 1 1 4 4 1 1 1 1 56 1 1 2 1 1 1 1 3 1 57 2 1 2 2 1 1 1 1 1 58 1 1 2 4 1 3 1 1 1 59 1 1 3 1 3 1 1 1 1 60 1 1 4 1 1 1 1 1 4 61 1 1 4 1 1 1 3 1 1 62 2 2 1 1 1 1 1 4 1 63 1 2 1 1 1 1 1 1 4 64 1 2 1 1 1 2 1 1 1 65 1 2 1 4 1 1 1 1 1 66 1 2 2 1 1 1 1 1 1 67 1 2 3 1 1 1 1 1 1 68 2 2 4 1 1 1 1 1 1 69 1 3 1 1 1 1 2 1 1 70 1 3 1 1 2 1 1 1 1 71 1 3 1 1 3 1 1 1 1 72 1 3 2 1 1 1 1 1 1 73 2 3 3 1 1 1 1 1 1 74 1 4 1 1 1 1 1 1 4 75 1 4 1 1 1 4 1 1 1 76 1 4 1 3 1 1 1 1 1 77 1 4 2 1 1 1 1 1 1 78 2 1 1 1 1 1 4 4 4 79 1 1 1 1 1 3 3 1 2 80 1 1 1 1 2 4 1 1 4 81 1 1 1 1 3 1 3 2 1 82 1 1 1 1 4 1 1 4 2 83 1 1 1 1 4 4 1 1 4 84 1 1 1 2 1 2 1 2 1 85 1 1 1 3 1 4 1 1 4 86 1 1 1 3 1 4 3 1 1 87 1 1 1 4 4 4 1 1 1 88 1 1 2 1 3 1 1 1 3 89 1 1 2 1 4 1 2 1 1 90 1 1 4 1 1 1 1 4 2 91 1 1 4 2 1 1 1 3 1 92 1 1 4 4 1 4 1 1 1 93 1 2 4 1 1 1 4 1 1 94 1 2 2 1 1 1 2 1 1 95 1 2 2 1 3 1 1 1 1 96 1 2 2 3 1 1 1 1 1 97 1 3 1 1 1 1 2 2 1 98 1 3 1 1 1 4 4 1 1 99 1 3 1 1 3 2 1 1 1 100 1 3 4 1 1 1 1 1 4 101 1 3 4 1 1 1 2 1 1 102 1 4 1 1 1 1 1 4 4 103 1 4 2 1 4 1 1 1 1 104 1 1 1 1 1 4 4 4 4 105 1 1 1 4 4 1 4 1 4 106 1 1 3 4 1 4 1 1 2 107 1 1 3 4 4 4 1 1 1 108 1 2 1 1 3 3 4 1 1 109 1 4 1 4 1 1 1 3 4 110 1 4 2 1 4 1 1 4 1

PAGE 100

100Table E-1. Headway statistics for 16 lead-trail vehicle combinations Vehicle in position 2 Trailing 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Leading 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 count 61 6 7 6 6 4 1 3 4 1 1 0 6 2 2 0 mean 3.171 4.217 5.060 6.516 3.836 6.413 5.050 8.157 3.644 4.440 7.920 4.457 6.238 6.130 stdev 0.844 0.744 0.801 0. 967 0.972 2.423 1.608 0.464 0.627 3.115 0.170 Vehicle in position 3 Trailing 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Leading 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 count 63 11 4 8 4 1 0 1 4 1 0 0 9 1 2 1 mean 2.512 3.482 4.344 5.212 2.875 3.605 5.830 2.923 3.800 4. 158 3.510 4.835 3.900 stdev 0.785 0.546 1.410 0.765 0.714 0.768 0.924 2.029 Vehicle in position 4 Trailing 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Leading 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 count 61 5 5 9 7 0 0 0 7 0 0 0 11 1 0 4 mean 2.360 2.922 3.097 4.670 2.880 4.120 3.842 3.360 5.739 stdev 0.631 0.240 0.619 0.955 0.637 1.298 1.054 0.349 Vehicle in position 5 Trailing 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Leading 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 count 58 5 5 13 6 0 1 0 5 1 1 0 11 1 0 3 mean 2.235 2.578 2.708 4.293 2.929 4.620 5.132 3.620 3.990 3.435 4.500 4.290 stdev 0.644 0.389 0.275 0.680 0.581 1.048 0.950 0.231 APPENDIX E HEADWAY STATISTICS

PAGE 101

101Table E-1. Continued Vehicle in position 6 Trailing 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Leading 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 count 61 6 4 12 9 1 0 0 5 0 2 1 6 0 1 2 mean 2.297 3.548 2.833 3.973 3.147 3.165 2.876 5. 195 4.240 5.331 4.000 5.345 stdev 0.509 0.673 0.383 0. 747 1.135 0.975 1.520 1.653 0.714 Vehicle in position 7 Trailing 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Leading 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 count 57 7 6 6 8 1 2 1 8 1 0 0 10 1 0 2 mean 2.080 2.626 2.649 4.288 2.691 3.070 3.040 5. 090 3.239 7.590 3.807 4.490 5.895 stdev 0.436 0.403 0.513 0. 647 0.686 0.566 0.826 1.575 1.407 Vehicle in position 8 Trailing 1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4 Leading 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 count 56 11 6 7 5 1 1 2 3 0 0 0 12 0 2 4 mean 2.133 2.533 2.701 3.982 2.676 3.010 6.970 3. 350 3.728 3.933 4.925 5.377 stdev 0.471 0.295 0.715 0. 549 0.438 0.269 0.652 1.116 3.231 0.900

PAGE 102

102 APPENDIX F CALIBRATION RESULTS AND MSE CALCULATIONS Table F-1. Simulation model calibration results Vehicle composition Field Simulation MSE calculation Case 1 2 3 4 5 6 7 8 2-8 5-8 2-8 5-8 2-8 5-8 1 1 1 1 1 1 1 1 1 2.36 2.18 2.39 2.05 0.001 0.018 2 1 1 1 1 1 1 1 2 2.46 2.64 2.50 2.25 0.002 0.155 3 1 1 1 1 1 1 1 3 2.66 2.59 2.60 2.41 0.003 0.031 4 1 1 1 1 1 1 1 4 2.80 2.69 2.75 2.67 0.002 0.000 5 1 1 1 1 1 1 2 1 2.61 2.56 2.57 2.38 0.001 0.034 6 1 1 1 1 1 1 3 1 2.97 2.68 2.74 2.66 0.050 0.000 7 1 1 1 1 1 1 4 1 2.92 3.19 3.01 3.14 0.009 0.003 8 1 1 1 1 1 2 1 1 2.47 2.44 2.58 2.40 0.012 0.002 9 1 1 1 1 1 3 1 1 2.75 2.94 2.76 2.70 0.000 0.061 10 1 1 1 1 1 4 1 1 3.13 3.33 3.03 3.18 0.009 0.022 11 1 1 1 1 2 1 1 1 2.57 2.44 2.60 2.41 0.000 0.001 12 1 1 1 1 4 1 1 1 2.81 3.04 3.06 3.22 0.059 0.034 13 1 1 1 2 1 1 1 1 2.42 2.29 2.60 2.24 0.033 0.003 14 1 1 1 4 1 1 1 1 2.94 2.86 3.09 2.82 0.023 0.002 15 1 1 2 1 1 1 1 1 2.77 2.40 2.62 2.09 0.022 0.092 16 1 1 3 1 1 1 1 1 2.52 2.09 2.81 2.13 0.085 0.002 17 1 1 4 1 1 1 1 1 3.03 2.39 3.12 2.20 0.007 0.034 18 1 2 1 1 1 1 1 1 2.70 2.38 2.62 2.06 0.007 0.100 19 1 3 1 1 1 1 1 1 2.71 2.24 2.85 2.15 0.020 0.007 20 1 4 1 1 1 1 1 1 2.89 2.11 3.15 2.21 0.071 0.010 21 2 1 1 1 1 1 1 1 2.50 2.18 2.64 2.10 0.019 0.007 22 3 1 1 1 1 1 1 1 2.70 2.28 2.87 2.15 0.030 0.016 23 4 1 1 1 1 1 1 1 2.81 2.18 3.21 2.20 0.159 0.000 24 1 1 1 1 1 1 2 2 2.69 2.70 2.69 2.57 0.000 0.016 25 1 1 1 1 1 1 3 2 3.21 3.33 2.85 2.87 0.129 0.206 26 1 1 1 1 1 1 3 4 2.89 3.42 3.03 3.18 0.021 0.057 27 1 1 1 1 1 1 4 4 3.03 3.42 3.24 3.54 0.042 0.013 28 1 1 1 1 1 2 3 1 3.35 3.97 2.91 2.96 0.196 1.024 29 1 1 1 1 1 2 4 1 2.97 3.57 3.15 3.39 0.032 0.033 30 1 1 1 1 1 2 1 4 3.31 3.51 2.91 2.97 0.156 0.288 31 1 1 1 1 1 3 2 1 2.75 2.66 2.95 3.04 0.040 0.142 32 1 1 1 1 1 4 2 1 3.36 4.30 3.22 3.51 0.019 0.612 33 1 1 1 1 2 1 2 1 3.00 2.96 2.78 2.73 0.050 0.052 34 1 1 1 1 2 1 1 4 2.93 3.19 2.91 2.96 0.001 0.049 35 1 1 1 1 2 2 1 1 2.89 3.02 2.77 2.72 0.015 0.092 36 1 1 1 1 3 1 2 1 3.05 3.05 2.96 3.04 0.008 0.000 37 1 1 1 1 3 1 4 1 3.26 3.83 3.29 3.63 0.001 0.038 38 1 1 1 1 3 3 1 1 2.97 3.50 3.07 3.24 0.010 0.070 39 1 1 1 1 4 1 2 1 2.93 3.10 3.25 3.57 0.100 0.215 40 1 1 1 1 4 1 1 2 2.86 3.08 3.17 3.42 0.094 0.120 41 1 1 1 2 1 1 2 1 3.30 2.77 2.79 2.57 0.255 0.042 42 1 1 1 2 1 1 4 1 3.12 2.91 3.17 3.24 0.002 0.111

PAGE 103

103 Table F-1. Continued Vehicle composition Field Simulation MSE calculation Case 1 2 3 4 5 6 7 8 2-8 5-8 2-8 5-8 2-8 5-8 43 1 1 1 2 1 1 1 4 2.72 2.50 2.93 2.81 0.044 0.097 44 1 1 1 2 3 1 1 1 2.84 2.54 2.93 2.82 0.010 0.084 45 1 1 1 4 1 1 2 1 2.69 2.81 3.26 3.12 0.324 0.096 46 1 1 1 4 1 1 3 1 3.85 3.70 3.40 3.36 0.201 0.114 47 1 1 1 4 1 1 4 1 3.41 2.96 3.55 3.60 0.020 0.410 48 1 1 1 4 1 1 1 3 3.43 3.44 3.25 3.11 0.032 0.108 49 1 1 1 4 1 4 1 1 3.32 3.66 3.54 3.61 0.051 0.003 50 1 1 2 1 1 1 3 1 2.49 2.44 2.93 2.64 0.190 0.039 51 1 1 2 4 1 1 1 1 2.86 2.66 3.24 2.73 0.141 0.004 52 1 1 4 1 1 1 1 2 2.97 2.65 3.24 2.42 0.075 0.053 53 1 1 4 1 1 3 1 1 3.67 2.63 3.42 2.74 0.059 0.012 54 1 1 4 1 4 1 1 1 3.59 3.08 3.58 3.00 0.000 0.005 55 1 1 4 4 1 1 1 1 3.65 2.83 3.57 2.62 0.006 0.044 56 1 2 1 1 1 1 3 1 2.93 3.23 2.95 2.63 0.000 0.363 57 1 2 2 1 1 1 1 1 2.81 2.28 2.81 2.08 0.000 0.042 58 1 2 4 1 3 1 1 1 3.53 3.45 3.57 2.70 0.002 0.576 59 1 3 1 3 1 1 1 1 2.89 2.29 3.15 2.39 0.067 0.010 60 1 4 1 1 1 1 1 4 3.09 2.25 3.37 2.59 0.080 0.116 61 1 4 1 1 1 3 1 1 2.96 2.66 3.47 2.74 0.260 0.008 62 2 1 1 1 1 1 4 1 3.40 3.79 3.21 3.06 0.036 0.525 63 2 1 1 1 1 1 1 4 2.91 2.74 2.95 2.63 0.002 0.012 64 2 1 1 1 2 1 1 1 2.40 2.48 2.83 2.41 0.182 0.006 65 2 1 4 1 1 1 1 1 2.96 2.32 3.27 2.16 0.093 0.027 66 2 2 1 1 1 1 1 1 2.61 1.89 2.83 2.08 0.047 0.037 67 2 3 1 1 1 1 1 1 3.08 2.09 3.00 2.10 0.006 0.000 68 2 4 1 1 1 1 1 1 3.05 2.37 3.28 2.16 0.054 0.044 69 3 1 1 1 1 2 1 1 2.74 2.88 3.06 2.48 0.103 0.162 70 3 1 1 2 1 1 1 1 2.86 1.92 3.06 2.27 0.039 0.123 71 3 1 1 3 1 1 1 1 2.65 1.92 3.17 2.39 0.280 0.223 72 3 2 1 1 1 1 1 1 2.82 2.18 3.07 2.13 0.064 0.002 73 3 3 1 1 1 1 1 1 3.42 2.35 3.17 2.09 0.062 0.069 74 4 1 1 1 1 1 1 4 3.50 3.51 3.44 2.60 0.004 0.824 75 4 1 1 1 4 1 1 1 3.08 3.19 3.68 3.02 0.363 0.032 76 4 1 3 1 1 1 1 1 2.93 1.69 3.51 2.13 0.335 0.195 77 4 2 1 1 1 1 1 1 3.16 2.13 3.39 2.16 0.053 0.001 78 1 1 1 1 1 4 4 4 4.36 5.62 3.73 4.41 0.403 1.462 79 1 1 1 1 3 3 1 2 3.76 4.06 3.19 3.46 0.316 0.363 80 1 1 1 2 4 1 1 4 3.40 4.01 3.43 3.71 0.001 0.092 81 1 1 1 3 1 3 2 1 2.97 2.72 3.28 3.31 0.097 0.347 82 1 1 1 4 1 1 4 2 3.09 3.00 3.65 3.81 0.319 0.647 83 1 1 1 4 4 1 1 4 3.21 3.55 3.78 4.02 0.327 0.216 84 1 1 2 1 2 1 2 1 3.28 3.03 2.98 2.72 0.091 0.092 85 1 1 3 1 4 1 1 4 2.81 2.80 3.57 3.45 0.574 0.434 86 1 1 3 1 4 3 1 1 3.00 2.97 3.63 3.58 0.399 0.372 87 1 1 4 4 4 1 1 1 3.43 2.94 4.03 3.42 0.364 0.227

PAGE 104

104 Table F-1. Continued Vehicle composition Field Simulation MSE calculation Case 1 2 3 4 5 6 7 8 2-8 5-8 2-8 5-8 2-8 5-8 88 1 2 1 3 1 1 1 3 2.84 2.44 3.13 2.70 0.088 0.067 89 1 2 1 4 1 2 1 1 3.09 2.49 3.44 3.06 0.120 0.330 90 1 4 1 1 1 1 4 2 3.56 2.92 3.72 3.19 0.027 0.075 91 1 4 2 1 1 1 3 1 3.30 2.57 3.68 2.75 0.141 0.033 92 1 4 4 1 4 1 1 1 3.06 2.46 4.06 2.90 0.995 0.190 93 2 4 1 1 1 4 1 1 4.37 3.39 3.73 2.96 0.415 0.182 94 2 2 1 1 1 2 1 1 3.38 2.10 3.00 2.36 0.145 0.066 95 2 2 1 3 1 1 1 1 3.83 2.46 3.16 2.37 0.451 0.008 96 2 2 3 1 1 1 1 1 2.78 2.32 3.15 2.08 0.140 0.059 97 3 1 1 1 1 2 2 1 2.88 2.57 3.25 2.81 0.139 0.057 98 3 1 1 1 4 4 1 1 3.65 3.61 3.82 3.81 0.030 0.043 99 3 1 1 3 2 1 1 1 3.38 2.97 3.35 2.71 0.001 0.071 100 3 4 1 1 1 1 1 4 3.21 2.34 3.65 2.55 0.195 0.044 101 3 4 1 1 1 2 1 1 3.19 2.43 3.60 2.48 0.171 0.003 102 4 1 1 1 1 1 4 4 3.59 3.42 3.91 3.45 0.106 0.001 103 4 2 1 4 1 1 1 1 3.95 2.83 3.86 2.59 0.007 0.055 104 1 1 1 1 4 4 4 4 4.07 4.82 4.21 5.24 0.019 0.172 105 1 1 4 4 1 4 1 4 3.66 3.28 4.27 3.83 0.375 0.307 106 1 3 4 1 4 1 1 2 4.03 3.04 3.96 3.15 0.006 0.012 107 1 3 4 4 4 1 1 1 3.58 3.04 4.27 3.29 0.471 0.059 108 2 1 1 3 3 4 1 1 3.76 3.80 3.78 3.80 0.000 0.000 109 4 1 4 1 1 1 3 4 3.30 2.12 4.23 3.10 0.868 0.972 110 4 2 1 4 1 1 4 1 4.58 3.87 4.30 3.37 0.074 0.247

PAGE 105

105 LIST OF REFERENCES 1. Highway Capacity Manual. Transpor tation Research Board (TRB), National Research Council, Washington, D.C., 2000. 2. Washburn, S.S., Courage, K.G., and MacKen zie, S. Development of a Red Light Violation Data Collection Tool Southeastern Transportation Center, Final Report. 2001. http://stc.utk.edu/htm/pdf%20files/RLR_TRB1.pdf 3. Highway Capacity Manual. Highway Res earch Board, Special Report 87, National Research Council, Publicati on 1328, Washington D.C., 1965. 4. Molina, CJ, Jr. Development of Passenger Car Equivalencies for Large Trucks at Signalized Intersections. Institute of Tran sportation Engineers, Journal vol. 57 Issue 11, pp. 33-37. November 1987. 5. Benekohal, R and Zhao, W. Delay-Base d truck Equivalencies at Signalized Intersections: Results and Field Data. Third International Symposium on Highway Capacity, Copenhagen, Denmark. June 1998. 6. Kockelman, K and Shabih, R. Effect of Light-Duty Trucks on the Capacity of Signalized Intersections. Jour nal of Transportation Engineer ing: American Society of Civil Engineers, Volume: 126 Issue: 6 pp. 506-512. November 2000. 7. Bonneson, JA and Messer, CJ. Phase Ca pacity Characteristics from Signalized Interchange and Intersection Approaches. Transportation Research Record: Journal Transportation Research Board. No. 1646. TRB, National Research Council. Washington D.C., pp. 96-105. 1998. 8. Bonneson, JA, Nevers, B., Zegeer, J., Nguye n, T. and Fong, T. Guidelines for Quantifying the Influence of Area Type and Other Factors on Saturation Flow Rate. Final Report #DO2319. Florida Department of Transporta tion. Tallahassee, FL. June 2005. 9. Perez-Cartagena, R. and Tarko, A. Calibra tion of Capacity Parameters for Signalized Intersections in Indiana. 84th Annual Meeting of the Tran sportation Research Board, January 9-13, 2005, Washington D.C. 10. Li, H. and Prevedouros, P. Detailed Obse rvations of Saturation Headway and StartUp Lost Times. Journal of the Tran sportation Research Board, TRR 1802, Washington, D.C., 2002. 11. Cohen, S. L. Application of Car-Followi ng Systems to Queue Discharge Problem at Signalized Intersections. Transportati on Research Record: Journal of the Transportation Research Board, TRR 1802, Washington, D.C., 2002. 12. Greenshields, B. D. Distance and Time Required to Overtake and Pass Vehicles. Highway Research Board Proceedings, Vol. 15, 1935, pp.332-342.

PAGE 106

106 13. Bonneson, James A. Modeling Queued Driv er Behavior at Signalized Junctions. Transportation Research Record: Journal Transportation Research Board. No. 1365. TRB, National Research Council. Washington D.C., pp. 99-107. 1992. 14. Briggs T. Time Headways on Crossing th e Stop Line after Queueing at Traffic Lights. Traffic Engineering, July & Control, May 1977, pp. 264-265. 15. Messer C. J. and Fambro D. B. Effects of Signal Phasing and Length of Left-Turn Bay on Capacity. In Transportation Resear ch Record 644, TRB National Research Council, Washington, D.C., 1977, pp. 95-101. 16. Buhr J.H., Whiston R.H., Brewer K.A., a nd Drew D.R. Traffic Characteristics for Implementation and Calibration of Freeway Merging Control Systems. In Highway Research Record 279, HRB, National Research Council, Washington, D.C., 1969, pp. 87-106. 17. Evans L. and Rothery R.W. Influence of Vehicle Size and Performance on Intersection Saturation Flow. Proc., 8th International Symposium on Transportation and Traffic Theory, University of Toronto Press, Toronto, Ontario, Canada, 1981, pp. 193-222. 18. George E.T. and Heroy F.M. Starting Res ponse of Traffic at Signalized Intersections. Traffic Engineering, July 1966, pp. 39-43. 19. Akelik R., Besley M. and Roper R. Fundame ntals Relationships for Traffic Flows at Signalised Intersections. Research Report ARR 340. ARRB Transport Research Ltd, Vermont South, Australia. 1999 20. Pipes L.A. An Operational Analysis of Tra ffic Dynamics, Journal of Applied Physics, Vol. 24, NO. 3, 1953, pp. 274-287. 21. May, A. Traffic Flow Fundamentals. Pub lished by Prentice-Hall, Inc., 1990, Chapter 6, pp. 160-191. 22. Long, Gary. Intervehicle Spacing and Qu eue Characteristics. Transportation Research Record: Journal Transportation Research Board. No. 1796. TRB, National Research Council. Washington D.C., pp. 86-96. 2002. 23. Florida Department of Transportation. ARTPLAN Level Of Service analysis software. http://www.dot.state.fl.us/planni ng/systems/sm/los/los_sw2.htm 24. Cohen, S. L. Application of Car-Fo llowing Systems in Microscopic Time-Scan Simulation Models. Journal of the Tr ansportation Research Board, TRR 1802, Washington, D.C., 2002. 25 StatSoft, Inc. (2006). STATISTI CA (data analysis software system), version 7.1 www.statsoft.com.

PAGE 107

107 BIOGRAPHICAL SKETCH Carlos Cruz-Casas is a 24-year-old graduate student at the University of Florida. He completed a Master of Engineering degree in tran sportation engineering. He received a Bachelor of Science in civil engineering degree from the University of Puerto RicoMayagez in May of 2005.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101114_AAAAIJ INGEST_TIME 2010-11-14T22:05:44Z PACKAGE UFE0021137_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 1051950 DFID F20101114_AACWMB ORIGIN DEPOSITOR PATH cruzcasas_c_Page_008.jp2 GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
f880fd38e8a09c6ede69abb8c62bb65b
SHA-1
34fed8ab6c3719285000a616d7eb559e60554305
44632 F20101114_AACWLN cruzcasas_c_Page_093.jpg
38322695444ed635a20246f6e863ec2b
cecd9806c0430e6504e2413bf0452414805bcbde
80978 F20101114_AACWKZ cruzcasas_c_Page_072.jpg
386ee7ef063fb78e1262b4e2d58e137f
711fa0d0aa8663686963c9de3956fe61f96c6923
1051984 F20101114_AACWMC cruzcasas_c_Page_009.jp2
6a058e62c773c48d419866b8ade157c2
c8129722de8d0c1632765a8e5e048c54551b8927
17976 F20101114_AACWLO cruzcasas_c_Page_096.jpg
98d0edcfe8584176e47db45a1837856a
41bf4c4c1738d42bb5085f3f984c77d9d7d281b8
98861 F20101114_AACWMD cruzcasas_c_Page_010.jp2
b212da532aed172a051db351e2314672
cfe3e58eb3ede12bea4ed2f529f2f1c57311f762
55835 F20101114_AACWLP cruzcasas_c_Page_097.jpg
b5e2997318bda5ce5787b10b03cd70dc
a04dc4016af812851c530f3de2538a81b93dea0c
68418 F20101114_AACWME cruzcasas_c_Page_011.jp2
39af311c5c15cebd263d31d6fd4549cd
d2c9243b5c9245c3d4678540d9a71c14f455d4cf
47194 F20101114_AACWLQ cruzcasas_c_Page_098.jpg
dc3085a4fa42d0d4221cc484648d3a52
68211672c2cdb54ef0fa27d8d8323c733283277b
110758 F20101114_AACWMF cruzcasas_c_Page_012.jp2
8fa3a2752c53fb494d00064b39aa8412
df90aa0e02bf346d385047ee14c9a734e76f0fb8
37277 F20101114_AACWLR cruzcasas_c_Page_100.jpg
a0e0894ea275cdd79e0a356df655418d
cb5ff7570341360b1db3d77c908231689a248217
116569 F20101114_AACWMG cruzcasas_c_Page_013.jp2
ec06639f0f064850d24780aec23aa2c1
258abc69f5c1e575b963470f6dc247e60bb5c51b
90387 F20101114_AACWLS cruzcasas_c_Page_102.jpg
6ff86786b1c7ffcf3402894b2b486a2c
d5ac3b69f2ae2cb875836a1567631ee3312328ea
107433 F20101114_AACWMH cruzcasas_c_Page_014.jp2
a2cf062aceb50851dbe4bd76652529f8
2a9c9d8590a138f0360edcbaeba70d07463f83c7
52555 F20101114_AACWLT cruzcasas_c_Page_104.jpg
18aaaf036b402c775781565bcd802642
9e0e7d4fad248dc719b16b934514e48e5ae1d9cb
127257 F20101114_AACWLU cruzcasas_c_Page_105.jpg
6fd268611931b840696ffcfa8b52dd17
847adecb84ece93928b6b21502b06a66f3bff2fe
17157 F20101114_AACWMI cruzcasas_c_Page_015.jp2
33222300593050ecab46d42e8c266c28
2add8044d34246dd78a80dde604345622f636761
21461 F20101114_AACWLV cruzcasas_c_Page_107.jpg
7e97de04ef301cddf7d02da9edc1b653
fd985d53a4ef85d81a6f3702705ffbc1cfd8da0c
111124 F20101114_AACWMJ cruzcasas_c_Page_016.jp2
6284dfbbacb9552523b69b68bacc172a
b8e343279bc8934df41ced700a17bd7501ca5e53
5688 F20101114_AACWLW cruzcasas_c_Page_002.jp2
9a24e90688931f6d84909122d119c7a4
82f0260d2aa7a8632612c80d7dc05da601ac8062
120616 F20101114_AACWMK cruzcasas_c_Page_019.jp2
df6fc2035147acf952c0cf2646f8f38b
89ab64c4eba744fdc44a9e658623c728d2430185
6084 F20101114_AACWLX cruzcasas_c_Page_003.jp2
1d1bc8997e900b17ab0b83504ab6d72c
4a6c7aef1ec4e69d18428f9f701441cc36d5b241
115908 F20101114_AACWML cruzcasas_c_Page_020.jp2
4c0ae027a5d4f7202b3d1a6dda1dee34
3f344d4559ceedfc1a406c564a83f51e9de1cf95
30508 F20101114_AACWLY cruzcasas_c_Page_004.jp2
39b77756e2349c64c7590bbecb33a828
557a46bf0673da5c6d8fc3109635b251c0a3531b
1051856 F20101114_AACWNA cruzcasas_c_Page_047.jp2
0a4dbdf66b93a9ce91c57b81109920f3
8feb8de3cadcc141c39bc8f0acb9603354feda29
116433 F20101114_AACWMM cruzcasas_c_Page_021.jp2
c9c61b0ee9ea8dd9e22c1669aed28c47
773a42b131674ad9774273d183a202f2710d28fa
714246 F20101114_AACWLZ cruzcasas_c_Page_006.jp2
2d1969afba015e84f00401503f11a64b
83aec85de0e65088c9042530e67140d9ddc13cad
90318 F20101114_AACWNB cruzcasas_c_Page_048.jp2
98f07527247972c4e07ee3ecb72dbd94
978a8bdac3bda751e1891c426679a5d3d67ae4c5
113678 F20101114_AACWMN cruzcasas_c_Page_022.jp2
09a98c0006038d771ea2b276fda54c04
bc1e58772800ce365fe68f51790b6b41d2b8bf75
105784 F20101114_AACWNC cruzcasas_c_Page_050.jp2
93243e2cb53c6305e2c9a0b577cd96da
d9a1668122a970d5e44ec2b8179488ce0a242d7d
99306 F20101114_AACWMO cruzcasas_c_Page_023.jp2
3b717f765c8405d38476862781541143
e5ce111d9db8556699506353d56e60d359f436f7
1051912 F20101114_AACWND cruzcasas_c_Page_052.jp2
de495e1574c1c0137d2c76bc6d1c6875
90ccd97883ac2f7055136a60e022d287ce3ffbae
78157 F20101114_AACWMP cruzcasas_c_Page_028.jp2
9feb31273b839ff4b21c2c7498c1ce56
dd6e325c9458c5bfd4baaaaddff9df7959107c95
115297 F20101114_AACWNE cruzcasas_c_Page_053.jp2
51d2ffd2db7f694629957f727335f68d
dcee6e687663a3cd5eed99b1715549f26f8b0d0c
101604 F20101114_AACWMQ cruzcasas_c_Page_029.jp2
0b1090b3d65bea400c19162169fb3d01
b9c28e633e4b2dee39d5ab2e0283deb486dba234
117628 F20101114_AACWNF cruzcasas_c_Page_054.jp2
17c4c234998414be6ae0ea01f5e32fd5
b8b3c593078beec8df6da8e136b8fda3970192ba
100654 F20101114_AACWMR cruzcasas_c_Page_030.jp2
425241973b564c0ceba65786c1c1c8dc
6f95f0272b653f2133e395e6a6faac5fc7533e2a
37006 F20101114_AACWNG cruzcasas_c_Page_055.jp2
11220b3ca0ca92ef18b001ece3db738e
3cdeec61279894c864f8badc70dd822566a17172
112696 F20101114_AACWMS cruzcasas_c_Page_032.jp2
9f45327700f3c50d34ce665c863d4b44
a824f96be91a858793a53a0c818a812171d2c81b
88318 F20101114_AACWNH cruzcasas_c_Page_057.jp2
23dcf5c333614ec9b9fe3c3a0d79472e
ba7f72e51d9615b485cefc25298e7ee2f489f2e2
110567 F20101114_AACWMT cruzcasas_c_Page_034.jp2
4f77fb930f58897a1b232adbadbade81
7fc9b43815c19fbff07fb35d2595de751a27ea0b
95975 F20101114_AACWNI cruzcasas_c_Page_060.jp2
a4736d028fd28cf489bb2e49f9d5f657
78a07c8c2cedfc5513a06d457267400aba26c4ef
98552 F20101114_AACWMU cruzcasas_c_Page_037.jp2
2604102d04aa130fe7937a5c4d5185c4
cd76084d06658775fc8343bafeb5cf8d5f9409e4
644886 F20101114_AACWNJ cruzcasas_c_Page_061.jp2
fc4d58c0462091264209ab9280b68a99
016e3ed705b9b40098d778126f5c2f6163a66218
1051958 F20101114_AACWMV cruzcasas_c_Page_039.jp2
3b8e5fae660838a73a168a26e4b33380
3c4cd0e5420720dee4a9fb8bfee66fabdc51cf40
95162 F20101114_AACWNK cruzcasas_c_Page_062.jp2
7afc8796a0781169fdb4d2711d622fb2
04fa8ae525510a5dcaa07c5266fbad8f79075c1e
1051985 F20101114_AACWMW cruzcasas_c_Page_041.jp2
19631799df4bf97ec1faeb83f7f5e579
14b4234083422529608e41c78f52e8cbf9c30dac
98468 F20101114_AACWOA cruzcasas_c_Page_082.jp2
35d6c41bd34a5f0636a38929a21d8a68
b2ecdde3b5f547a892cda345c374ffe5daa64428
99196 F20101114_AACWNL cruzcasas_c_Page_063.jp2
35862d80d2e00408c83da3de1b3a89ad
7adb2333c864e27d485530a0e66e1fffe72fbf41
1051978 F20101114_AACWMX cruzcasas_c_Page_043.jp2
5a315299da844a622cdd3c7fc17e6d27
3f0490cd62a0545b678398db9c0c72e2ae2ed41a
1051980 F20101114_AACWNM cruzcasas_c_Page_064.jp2
114051d9014ecb55acb271480f4cd0b9
548cb4e109b7da4e0b84593e1e3585533d7884e7
1036280 F20101114_AACWMY cruzcasas_c_Page_045.jp2
16e5aa77cdcc3a83501ad5bf90952d0d
b51310db93351734475bb8750396d961091cbe25
41205 F20101114_AACWOB cruzcasas_c_Page_083.jp2
40315ad79c67becc8bd917ac9ff24bf8
243b7de75e58d5f781bad16d7f912e7326000cd9
109648 F20101114_AACWNN cruzcasas_c_Page_065.jp2
9e9334461c95e231c517bc30fa9064ea
67a08b7dca3d3d13ea8e522b4ace6eb4a88b7706
1051943 F20101114_AACWMZ cruzcasas_c_Page_046.jp2
36865b4bafbda796040ba5156ecee22b
d7c727a6f75e142ddbe5db5879f14bc8e8c4722d
117988 F20101114_AACWOC cruzcasas_c_Page_084.jp2
e3a7b13dc055b45f56f69fdc4cd9f2fb
ea0698e065b854c9ad8c6439067507eed9a9e455
37193 F20101114_AACWNO cruzcasas_c_Page_066.jp2
6a7e2ffc6183496078ffbabd1f37e7e8
59dc9ca711b262d63fe93edb530ac0b6e9046fd3
117061 F20101114_AACWOD cruzcasas_c_Page_085.jp2
b892ac4264289b1fc2005489759f58ba
c6e2c055daff9c617ce495dffc65cedba58c7e2e
998519 F20101114_AACWNP cruzcasas_c_Page_067.jp2
eba9349e68110e84f7ea7f59bc589579
e15f98454d22712d506739714986f15c4715e1ad
111037 F20101114_AACWOE cruzcasas_c_Page_086.jp2
cc25ef90a03a079ace5d73e5ed8e42f3
b3c36b8514808994cbc95e5560e03c07cc430004
1051961 F20101114_AACWNQ cruzcasas_c_Page_068.jp2
54656a2e5ff0feb0fc5fb6a474c2dd14
88475d9702e21a65d8ee8037ceb85ab1a60d9182
115405 F20101114_AACWOF cruzcasas_c_Page_087.jp2
5416fe6917ec54db4a3c0aa3be95a5cb
5ef42f3e58f2f1d446dce32d9ad1fd81afc4cea8
117595 F20101114_AACWNR cruzcasas_c_Page_069.jp2
f81f3e642ed130457fd7371248b86b99
adc3724eaf161daf0fdab689dccbff420d7a34be
118756 F20101114_AACWOG cruzcasas_c_Page_088.jp2
5d9aea78d1f45940bd64bcb6a0b46637
3e528e22a54b6b352bc65141109ff9ecf592d10d
858676 F20101114_AACWNS cruzcasas_c_Page_072.jp2
9e52a204528c0fe2494e85f15843517a
044fccd4cb1bd684cfb7e1b4fb2e49ebc6562813
390790 F20101114_AACWOH cruzcasas_c_Page_090.jp2
a47855d32624a591e38f49f5316b2ba7
f5cb014777283c9c6d4f3462da7b58e5d2e969e1
77144 F20101114_AACWNT cruzcasas_c_Page_073.jp2
ba766a15b20118bf6853d9c27a7b06bd
3251d6b66be5a85171c8327965ee042ac9bc366c
1051982 F20101114_AACWOI cruzcasas_c_Page_091.jp2
31047cbe22b96f3c4d02786847c7695d
93ec2860ba111fbc7659aa409edf0897fb36e324
758073 F20101114_AACWNU cruzcasas_c_Page_074.jp2
1bfa0683ad46b2b13b5f659acade0640
5a8cc20711cca010ec8d4fa246cc8570df7fb260
818654 F20101114_AACWOJ cruzcasas_c_Page_094.jp2
926ce2ba0c10983e7633d5eea0ae84b0
3edf0dd42e8cabeff4a8a6e7574605dd6d1c6d0f
103909 F20101114_AACWNV cruzcasas_c_Page_076.jp2
d5aff29a8326ccebbac42b979b72d193
6bb60d0e2640a3b6d6bb24c05fac0ba3484367ff
36053 F20101114_AACWOK cruzcasas_c_Page_098.jp2
de6cb1cc8a1a0fdfb2a5711e0e04e97e
01ecd7cc75333f3a2bcf772eff33480286af748b
93959 F20101114_AACWNW cruzcasas_c_Page_077.jp2
58742ce1fa7e378a3fa071611b052439
63bd1b2ae4835a6c07a19ec1f57a35f1eac6e91e
37606 F20101114_AACWOL cruzcasas_c_Page_099.jp2
da8ca796b56353b15a2a90d044f375b9
b31c486ff67e65c98d6338b0d50873b9da3c3b82
89139 F20101114_AACWNX cruzcasas_c_Page_079.jp2
58dfaf104471a474d0fa0371cdb54e31
53a372b5794ba33808c0196790da57750c52af62
25271604 F20101114_AACWPA cruzcasas_c_Page_018.tif
cfb9a3931ded8e2d18d6eeadaff25875
81d3a1e0849627e7c2f069f0d5fea6121de534b2
39363 F20101114_AACWOM cruzcasas_c_Page_101.jp2
a673f35cc3371d47ca42269c9965f65e
caa98bed23bb5049f07cecf17e05cdbf38ab95d7
79093 F20101114_AACWNY cruzcasas_c_Page_080.jp2
4b8db24a1b3fc028c8cf00de54b9e23b
b61da59822502c10fe84dcd1ea9c2ed8072c1b85
1053954 F20101114_AACWPB cruzcasas_c_Page_019.tif
994d508efd711507b773627d8bb88382
ad1c2374ed1d694640a134a403ac41bb16959773
85573 F20101114_AACWON cruzcasas_c_Page_103.jp2
a05d7989f25ca44b39a4fe8411aec304
edd2e374302902ebb542a7c19fd2cca5ee89e985
92546 F20101114_AACWNZ cruzcasas_c_Page_081.jp2
e137c0a0e9685acc0b335904ecd6dcb4
4775995a676f36e24209522f00658645f27ad329
1051962 F20101114_AACWOO cruzcasas_c_Page_106.jp2
30f6c8db92d6bd356e31aba514b0782b
744bf3e5fe357b84930e0aa0d1af04b075740453
F20101114_AACWPC cruzcasas_c_Page_020.tif
9effa323cce0d9e3ce4ba5a268b889e3
73842fb2d5000a81339cf8f0291862589dfe2f71
F20101114_AACWOP cruzcasas_c_Page_002.tif
c8cb41ae77c819d8a06ca561600b2f5f
d41edd88fe048f6abdb06df990041d4be7028019
F20101114_AACWPD cruzcasas_c_Page_021.tif
21641a22cbdd45e85fa76ffb92b5c318
c4d4849a0fa28d424dc9a5b037f5aeb64df19179
F20101114_AACWOQ cruzcasas_c_Page_003.tif
f3ba768c18007e527688a00ff15cd0f9
b2ebce9cffc2dfd651eaaaebc76702ae24e8c859
F20101114_AACWPE cruzcasas_c_Page_022.tif
4466b8ccfe478cdef9c2c49103260b6d
6b67ae976f80a05eb259293ccb8856b15ff6819a
F20101114_AACWOR cruzcasas_c_Page_004.tif
bfa353968bb32a8fba00edf37ce5231f
6b6457c9f384becf955171154c4a7f41f747fd0a
F20101114_AACWPF cruzcasas_c_Page_023.tif
318a43fd78f6f5b77b1b43ab0c54cd34
15ab6b6102bb5448913bf5d44c59764d0f722c87
F20101114_AACWOS cruzcasas_c_Page_008.tif
ab60b1b5a2ad71f46e5ab8840592135e
06f16c1c2ad56456fbd74c7b34a79fca782b8473
F20101114_AACWPG cruzcasas_c_Page_025.tif
977145652028db9e6dcdd9bcabcd7271
c69bef24bea63ba4423fa0e04f26e9a1214ee084
F20101114_AACWOT cruzcasas_c_Page_009.tif
359ffa6d2a52740b4f269827f84c9756
93e517d692220283ffddb56a13b6e8c57a082313
F20101114_AACWPH cruzcasas_c_Page_026.tif
53e506a9b5fc215697a8f3a8ca2bf246
21b9cb17807ffad4ba1e0a01772229aa8a378444
F20101114_AACWOU cruzcasas_c_Page_010.tif
e3ff9a46fe3db45c06c78e4a968f047c
8d95d75ef3d1280949c19ae7a80faf5242e8eb37
F20101114_AACWPI cruzcasas_c_Page_027.tif
0f9335e72009c9b449ff30e4d76e3d1c
20eba6405918e555f850d4213fedaa93177edc04
F20101114_AACWOV cruzcasas_c_Page_011.tif
42984a5a35f14cb47d8b6afa1871fd17
09b0b8e492d06507290fc010c74e0f8d8fc535fa
F20101114_AACWPJ cruzcasas_c_Page_029.tif
7df73cb0d73760e76d4bd7cbcd917189
9d61af3d49debac44614dd50d8d6856b07e8e38e
F20101114_AACWOW cruzcasas_c_Page_013.tif
1aad7d8900388446c5c76ee46b68c582
1f142aa2682cec861e230a520076d30ce46a60b5
F20101114_AACWPK cruzcasas_c_Page_030.tif
5f10380a782abe623af6158fdff1adae
afe8e910851c9a6687272e4147cae1bc4d9544df
F20101114_AACWOX cruzcasas_c_Page_014.tif
f529d60371d483bafb4ab949e4568d04
f50197e32c342f4a70ab08c77ab713949c8e012d
F20101114_AACWQA cruzcasas_c_Page_050.tif
c41a1a49e79a0fa21409140649b5a702
8940a67cd454b15606771aa0369fce3118de664f
F20101114_AACWPL cruzcasas_c_Page_031.tif
99fd21408162fd4b84ecd53eb9968aee
72d31589f3cbb59568a2d6e9a304b99e80aaf1de
F20101114_AACWOY cruzcasas_c_Page_015.tif
e7b632a266b9b8579f8af526324ddc46
d836a4371a1dd02b81dd2c341fa3a819aca257c3
F20101114_AACWQB cruzcasas_c_Page_052.tif
ac5885e1a3f4a6d2b7b9141ff08b6dee
eec2986c890744c7ff9263a185ac3fdeaee502c1
F20101114_AACWPM cruzcasas_c_Page_032.tif
5fbc502e623e02c8763483644a5ab01b
7e53244ac06a023a30e5e30c46839793ff041bc5
F20101114_AACWOZ cruzcasas_c_Page_017.tif
24453aadd3c20309c95052ce2fca5470
cb06b5005dc3029c52b130149652a4bbd62a7735
F20101114_AACWQC cruzcasas_c_Page_054.tif
714b5c1dcdb04b3a401165651aac29dd
a5dd2e50681e3b2d2cad70033963f0d4419acb50
F20101114_AACWPN cruzcasas_c_Page_033.tif
f079dfe26624d870bcc460fae5194de5
a7f61d04bced31b2382d2dcd33de5cd26a033611
F20101114_AACWPO cruzcasas_c_Page_034.tif
3add3c7ca34fc0af11d0f8a30fb585a9
0cb7f9360b4a2f78a851c7f77d420340a17cb4b8
F20101114_AACWQD cruzcasas_c_Page_055.tif
c90941bbcd448a06d9bc01df2e20542e
f4a7f6af0bb62b6f5618b43493e154dad99e2dc3
F20101114_AACWPP cruzcasas_c_Page_035.tif
ac08891cd88b577a971b195f8c75472f
5ac3ba401ad09aa6a3f36b02a710dc17a86f2dad
F20101114_AACWQE cruzcasas_c_Page_056.tif
046620fb9a216a0e48b6aa60ae8f53db
9ddbd4932a00a980839714e08d247a028b92f460
F20101114_AACWPQ cruzcasas_c_Page_036.tif
cdeba25576b747ef5ff5a606561888d2
6530fe6fab1b57e83faad6b92ade5f957bcfac69
F20101114_AACWQF cruzcasas_c_Page_060.tif
22c72c5a7daafccb02b2d648d41e2884
83d87fd92653eb02c3e8c19c51807f3a1f6c6a70
F20101114_AACWPR cruzcasas_c_Page_037.tif
e0c2767302d2df9ab9cf19c66cb6a42f
da2eda17ef750eb305522d9881f47f0fcba9ea17
F20101114_AACWQG cruzcasas_c_Page_062.tif
7a2986bfab7516ffd6144cfe11362617
d355fcfc67739cfa5831d7e4a94edeb6a12b5fd3
F20101114_AACWPS cruzcasas_c_Page_038.tif
ea79be4bde786f610a0aa36f184fb11e
cfec5feab2a2766bf196a9bac98efd4889a721af
F20101114_AACWQH cruzcasas_c_Page_063.tif
5b532de0e80ac4612e6e39d262826b3a
e2a0300c61d31f2869be60807a6c709041af8ffe
F20101114_AACWPT cruzcasas_c_Page_042.tif
92d321c7d9dcc1484293ea9fc99e761e
36008ee9b775f41e6724a3ae3f86430456d460e3
F20101114_AACWQI cruzcasas_c_Page_065.tif
9b49ffb8968891369473a7f456855766
312dda4733018bf7958261942b4d405c1629c383
F20101114_AACWPU cruzcasas_c_Page_043.tif
722f3b7e37808a534ac0268793e5dd55
262f7ba7fee2dde2fa0379c546236aa206d748a8
F20101114_AACWQJ cruzcasas_c_Page_066.tif
9ddb958dafba62e339b27f7f73f16bd6
2141760c48bb434ad854249dfd875accb8fb30a5
F20101114_AACWPV cruzcasas_c_Page_044.tif
3f38fc543c00370c15aba76c792a0275
d42e75e0b13c035818544cc770a545b5bc48c929
8423998 F20101114_AACWQK cruzcasas_c_Page_067.tif
89aa32a89f1e0faa585535dfe3dbea61
c9c43d6f474ecd6c658dafa85007bc447b82503c
F20101114_AACWPW cruzcasas_c_Page_045.tif
2314b0ac7aeb87087defc7694d17bbc0
6a9c9d570ded6cc237c9fd22d665e232e8f4a138
F20101114_AACWRA cruzcasas_c_Page_088.tif
c526544af86d80e0b480de62071abcb6
00ea0afc6faed17dd992b34a85aee7e5dde6f76f
F20101114_AACWQL cruzcasas_c_Page_068.tif
2123294040f9ca552b91bd5d598b5145
ace708a35c7b120d62a71d0345be56241ed05f5e
F20101114_AACWPX cruzcasas_c_Page_046.tif
9e90acb7e0a8cf736e58f076fcf50cf1
571775cbc649e716a167f5856516619b0dcde848
F20101114_AACWRB cruzcasas_c_Page_089.tif
c8a033fddf48072ef74eb4d7921979bc
abf39139686c0e5f64226f344e61562b3ece5eea
F20101114_AACWQM cruzcasas_c_Page_069.tif
6b7c0ec0879a62e56623862a25e087e0
0bf7bb323669909e7949ff32170584ff7edbd40d
F20101114_AACWPY cruzcasas_c_Page_047.tif
82788e2de89aa4d04407860e4be5fa39
8f2aca2a17749354514af290841cd044a33c7da6
F20101114_AACWRC cruzcasas_c_Page_090.tif
7f22411bce4cde07fb66ae40f80eb6d0
3c19e9f263eb5058d3c80ed93b752ad1d43c7a81
F20101114_AACWQN cruzcasas_c_Page_072.tif
91774b20052ebf3c6ad7a57fde0e8d4c
1fc244ddeba4b364023cf03854fb4636f363b078
F20101114_AACWPZ cruzcasas_c_Page_048.tif
b6f8fc625dd9df37e27634bcede69777
18830e0dd8c8ff211e69f9512eb30c222e37ec5e
F20101114_AACWRD cruzcasas_c_Page_091.tif
abbea63f5fed173585a691ce67004c01
e2a716644f69259b0ddcd29fb92fa6c6bc2818e6
F20101114_AACWQO cruzcasas_c_Page_073.tif
8a3b3b812ef3cc344756e5cd38cd6233
a786564e09719dadca9bff999e66b6eaa58c357e
F20101114_AACWQP cruzcasas_c_Page_075.tif
74f89bf550315c5ea34553c0eab1ece3
4643a5c816fe16799cdec7355a073298e334d2d8
F20101114_AACWRE cruzcasas_c_Page_092.tif
7056855ec916fefe88cc5c98fb98a8ca
f290919ddcf1c1cfc2038f9b72d6a91a8b1c7b01
F20101114_AACWQQ cruzcasas_c_Page_077.tif
be5b64ba8e8e8ae66d12f0e4ae138008
1be55c33eab280cc109b942f40fcf6bae441875d
F20101114_AACWQR cruzcasas_c_Page_078.tif
ce06615c25b51e4751dee126381d343e
682a2fa5f1ba0dee171e620f83732242d66982be
F20101114_AACWRF cruzcasas_c_Page_093.tif
56f99fe9aa5580940623b361d0033d34
ee4fba2b2f5ed71e1a54917c1dba74c951c246de
F20101114_AACWQS cruzcasas_c_Page_079.tif
5bdfff7c3ba054b67e24db1122dd77ed
33a8e53a9666cfff5e9e309af00ab758a3fd6dd1
F20101114_AACWRG cruzcasas_c_Page_094.tif
8c8aa9395ca52ba09a530378c5cb1cb9
3bcb289995456dace48077c22ca8614072b4c8ce
F20101114_AACWQT cruzcasas_c_Page_081.tif
1a36b156b0754b53a2a6722809e6c4a3
e15491b0fbd3ba2f5afba59c2d77f3d3cac4b35e
F20101114_AACWRH cruzcasas_c_Page_095.tif
e9be73b2b854ce27afcc1a2a4500cddd
9df84114c9ac6088ad45ba28c6a4f88051e994fe
F20101114_AACWQU cruzcasas_c_Page_082.tif
73e4539d71811906bef81f1d8290cb34
de9e34af3db04fc904d0964263eca743500e7aa4
F20101114_AACWRI cruzcasas_c_Page_097.tif
b6aaf24e50cf1cdf5d36dc2b923658d3
d8d2f4318706f52907f51b2a099c5ee11ed82ecc
F20101114_AACWQV cruzcasas_c_Page_083.tif
2a92d387f52891d1fb2a21dee070a9ab
6d17b5fe2ab8ca984b2b0d8266cb918b3631a786
F20101114_AACWRJ cruzcasas_c_Page_098.tif
74668e4ebebb20b93fbad4bad4f5535b
908fa13b865277d2aac1b56342053fc206f91562
F20101114_AACWQW cruzcasas_c_Page_084.tif
7019ed09b8bd667dbda15a48878bc908
1adc9f2537dea862bcfba4491a618e33366c766d
F20101114_AACWRK cruzcasas_c_Page_099.tif
984eac109b7834bd599883fb505a6ddc
62be33aa6231ac4487fe4661de3f122c1c98f079
F20101114_AACWQX cruzcasas_c_Page_085.tif
45f4c7848add2ac149709b64d57bba4a
5704369d866f4da9dae9ec02a0cf482052978402
54599 F20101114_AACWSA cruzcasas_c_Page_021.pro
9ff0586db08f58bd2474d49c9d1023a9
279f76bc7e33ace3271d8c536c2057826c49d887
1054428 F20101114_AACWRL cruzcasas_c_Page_100.tif
218c9971139669afbe411aac728bedb4
2cb8536d37e09fbf88f03b44f46b4f7610aa2f7e
F20101114_AACWQY cruzcasas_c_Page_086.tif
519fba0c95eaebb5e5a5367aaefafc7e
ce37601b2f169d89fa8e6782e0ee2e007b73c990
53269 F20101114_AACWSB cruzcasas_c_Page_022.pro
986b08988e143b61308c987792b41fe1
84997a458a73afb72fd68eb3ca954c906037c4f3
F20101114_AACWRM cruzcasas_c_Page_101.tif
02ca77944fc0ea299e301d8b5db3a3c8
f5cc77b8ab1f5b0798c69bd0fae3a957cf8d6158
F20101114_AACWQZ cruzcasas_c_Page_087.tif
cb189e1c9ee9dcc9b32b5e4a11682b6f
1fde4a612bddc702f6e9ff7606d021a02d9158b0
44978 F20101114_AACWSC cruzcasas_c_Page_023.pro
f192cd65ef7db7ce27d69d1502269a1e
26b61161672bdc436c916219e318a5b1af7de950
F20101114_AACWRN cruzcasas_c_Page_102.tif
c41a69aa06102cd35ac94adae41aadb3
866be6227bd09929a82a8aa358c0e6c5c69983c3
54980 F20101114_AACWSD cruzcasas_c_Page_024.pro
8f5674d83150a40432236a028f4f311a
9920f74ee06e599d1101e2d2ad1e792f0d52cce7
F20101114_AACWRO cruzcasas_c_Page_103.tif
cf29b69a4c4355cba427736a8adecd14
9af3586b3388d72f53d9549ed601898dc12441c8
45083 F20101114_AACWSE cruzcasas_c_Page_026.pro
ce35689891c4ca03b21be1ad8eab3dd1
b4f186ef677e1abdb08aa62bcaae43ff05229789
F20101114_AACWRP cruzcasas_c_Page_106.tif
5115c1c87c52c395841a3bdc69095cc7
dab6d69259c2389291ca147074a302ca9300d7d0
8127 F20101114_AACWRQ cruzcasas_c_Page_001.pro
171efd1598858b97e16ce101ad8ac3dc
2014ea2f9aaca1f6d45fffdb7bcaa2c5f4844de9
36210 F20101114_AACWSF cruzcasas_c_Page_028.pro
6dbdb7bd86955f6f32b9d039bb2c681d
f03cfa7ef93e877fce1e833478d111f096f65a54
989 F20101114_AACWRR cruzcasas_c_Page_002.pro
f71f0161b7d69ac424a89485ade9e19e
5242e832eabbb6ea35e20665c0271ac987ff744f
48261 F20101114_AACWSG cruzcasas_c_Page_029.pro
01edfeddf15ec004d824a6681ec236d2
95c605bcfa0ef0365ef6cb81564ae41c35d01d1d
1327 F20101114_AACWRS cruzcasas_c_Page_003.pro
9388a13febf8d69c434f2caea741f637
bb43c5d490eb48834bf61c6780a8737de835b3d5
55708 F20101114_AACWSH cruzcasas_c_Page_031.pro
3a9af18d29fb72867aa2f09f048ecbc6
561e7e6b6d9774b03d3aa976923c78dca786495b
67418 F20101114_AACWRT cruzcasas_c_Page_005.pro
dc29981d378b89d3d95d6ab35035112f
0664d6a95c0c45b1ab09c9f7e3e8865df0e597db
56578 F20101114_AACWSI cruzcasas_c_Page_032.pro
3f5010a892980758ff0d3d914b0381db
f0989fb106a3616f855cfc145462174947ce46b7
16633 F20101114_AACWRU cruzcasas_c_Page_006.pro
030fc14aa99cae38116255d6f87cc41f
101fabf96b6c7fe4b4f5fd5b0802fce13c7e9469
5491 F20101114_AACWSJ cruzcasas_c_Page_033.pro
ba6353daaabdbb2fb78ec9744509b617
f14a824cf9cbcf3626fcba2ab27a597028ba9f4e
52330 F20101114_AACWRV cruzcasas_c_Page_012.pro
ba5682342cd03ee8ed344b090e744500
9ae2266d6bf4dab0c8d116bf43945db128de49c0
51572 F20101114_AACWSK cruzcasas_c_Page_034.pro
c9dd951fc7d4554e8e80687ba9638449
1e69f36c533c1e0095729464d47d0982215ed8ea
54790 F20101114_AACWRW cruzcasas_c_Page_013.pro
c12ceebf21599ee2c04e25dcdfc81a96
621565fd3f762e07adefbc2ce6d8a4be63b06235
15870 F20101114_AACWTA cruzcasas_c_Page_055.pro
4c7b35d2bcd22f808c113993f8495219
b29b6b561f5f99de4b9be22f7ed34a43c784bc3c
49142 F20101114_AACWSL cruzcasas_c_Page_035.pro
84639c5a33bc7933fc795fed74af1d4c
f36d141a0878bc1988120ddf46896e4a8035f9f1
6322 F20101114_AACWRX cruzcasas_c_Page_015.pro
99bcb8d3ca50abd35de4543fffbb6aff
fd216dcc563faabb67d2a13ce47f388623a16c05
49914 F20101114_AACWTB cruzcasas_c_Page_056.pro
ce310f2329ff2ab582ab145a6bc2f6e2
cafb3809e263439d860ecd6a047398705fa71063
57379 F20101114_AACWSM cruzcasas_c_Page_036.pro
c05bf8f79c999c997d1a4a934814fe3d
e27dbb0f1d89b604b6fe399d4787179f8dff6246
51971 F20101114_AACWRY cruzcasas_c_Page_016.pro
a5b46a09fac356ce2f5cee0140300c15
32c7f658f29b0e35d3ad7bcb6cd138f3f7d26803
34271 F20101114_AACWTC cruzcasas_c_Page_059.pro
2949e35016c8f21b1346b45a3ee64b56
e66df389e969ac6ffa75d39c444d1255c1dbc7f2
47671 F20101114_AACWSN cruzcasas_c_Page_037.pro
4fa080231658b5b641c69fd35e0d00fa
0fc5bb1e83b33bc98cc886790a2441eb21cb8975
42505 F20101114_AACWRZ cruzcasas_c_Page_017.pro
3bb64a6a6a4362bfdd3704a12367f672
18789a8885723777c978531235645a285dc76da3
46014 F20101114_AACWTD cruzcasas_c_Page_060.pro
85fea762633cf9ee858ceb4eec5f1230
98910a8bebbfc1d4350256f6b2d43e7401794f40
5742 F20101114_AACWSO cruzcasas_c_Page_039.pro
b9fb3adf72dba485c183cccfbcacefbb
7f9ba6f29d8272628a71244b4b8c036f6be73347
46104 F20101114_AACWTE cruzcasas_c_Page_063.pro
7ccdb1a08f11e82865e405a39fb790b9
aed155af10a94b2a858221daa450214d9725bbd1
2912 F20101114_AACWSP cruzcasas_c_Page_041.pro
0ac13351a77eed6184dc9dda7d6dfb32
399b6b29812d337c44301781939006bf2513800c
54932 F20101114_AACWTF cruzcasas_c_Page_064.pro
9807e79daed0770ac16818ed499fe73d
e13cb675e6c1cfd8720246223659837daa9ebaa9
4013 F20101114_AACWSQ cruzcasas_c_Page_042.pro
9e0ff26bd06498e03446fee3a5190432
85c337e90e851e2fc071ff7f2061d6690b1828e8
12483 F20101114_AACWSR cruzcasas_c_Page_043.pro
ecd895bc99fe990439d938f6e448b954
49d08bfea03230de785f177e445e5a61033f3161
52892 F20101114_AACWTG cruzcasas_c_Page_065.pro
c377881fcb2234e1fc13a8923dcde843
25422f54cafc434f51257058558822b069bbf630
8643 F20101114_AACWSS cruzcasas_c_Page_045.pro
707b8103e1dd65128efd82ee2b7ff316
2d3e75bd8e70af34683cb2f09789031816d89103
16562 F20101114_AACWTH cruzcasas_c_Page_066.pro
7515c86c926b7d6eb6d06c12ea14ddec
791a6d4a3c310285de65f4e55c7fb7292421ee9c
26760 F20101114_AACWST cruzcasas_c_Page_046.pro
5eaa2eee7d77bc6220914afd06b7a119
6fb6fe950a8feaa7b2cef4c0cacd85e90acb8bc9
43520 F20101114_AACWTI cruzcasas_c_Page_067.pro
59aec8c971cc3960a2edfdede5a06eee
631166d8b8ef55bef0446b50e70620a43da65582
22346 F20101114_AACWSU cruzcasas_c_Page_047.pro
39f0712e785f9e96455421dac102a353
3aa29d3fe2285153a7cff1fff3a1ce03f61997a4
51173 F20101114_AACWTJ cruzcasas_c_Page_071.pro
f6fe6e2959130f4e470509465d082a69
0bc1389d9309518d8ae3fe6357532bb3ad9b6088
47783 F20101114_AACWSV cruzcasas_c_Page_048.pro
97bdf3562d4b5e05599fa727c8aa38e5
79ad984f027e76f8e1dd5f28f30e7799041266b3
38219 F20101114_AACWTK cruzcasas_c_Page_073.pro
afaf496e2c3b1f8af8e7185bfb96b8d7
30d0182397a62531d574f492b0a6a95f20dd7500
50534 F20101114_AACWSW cruzcasas_c_Page_049.pro
dfd3d6889c1df82561241ebe3b949581
9dcd0e36a006bd68e98bda384dc47a10a0d24b6b
26346 F20101114_AACWTL cruzcasas_c_Page_074.pro
a9f6bcd4c6bc12acb7b49dd0c2ff1078
c91521637ed3e9b061be81fe676be2c6e3f39f48
48493 F20101114_AACWSX cruzcasas_c_Page_050.pro
8b31c6510bb6c0954363f28c7c94b8e5
6cee744162076e402199607ce9ed93bab6301f6d
4163 F20101114_AACWUA cruzcasas_c_Page_097.pro
049d0be8d532f0141af0ee12e5da4844
e1bb56e61e5eef5828299dd151e6da69208f340e
49123 F20101114_AACWTM cruzcasas_c_Page_075.pro
2f81cc790188c8208f8b8681513c1671
93ceff337e9b798ba1a7bd17490a6eadf562749d
54739 F20101114_AACWSY cruzcasas_c_Page_053.pro
c35e17c1e61b9615f6e4694263252c9a
2b25ff7ca6e33212bb50a02ef57c75da8054dec3
38064 F20101114_AACWUB cruzcasas_c_Page_099.pro
b3c291cdd6aa70612d99ffb88c54b88e
1469641b2202082b109b83dce3ed312e72ae4728
41401 F20101114_AACWTN cruzcasas_c_Page_079.pro
aff7b38fb3297bce0b2cf97760b49764
4a4412fb585e7087be8fbef85ea594820cedc0dc
55294 F20101114_AACWSZ cruzcasas_c_Page_054.pro
9a059e7a4f0fd2ae2415a9844db03cbe
876312e0b9c8a342e44062d8a83c1e550c025a59
23269 F20101114_AACWUC cruzcasas_c_Page_101.pro
8d98a4269dc29eaf0a5d3eeec499395d
78745921e0ad23e92eb75e60f87316ed066ecad3
36430 F20101114_AACWTO cruzcasas_c_Page_080.pro
b477980135234e3da09bb5539fd38f81
b58fa62c95ca52a7b815a2ee6f1b7657aca1735e
52402 F20101114_AACWUD cruzcasas_c_Page_102.pro
a8a8beaec21dd47869c741f28bebde7c
11d07527f201b821b54182a2cd738fd3de1569c2
F20101114_AACWTP cruzcasas_c_Page_081.pro
6b1ac4a92c452f8baee189a1599e6ab8
460bb602fefa0d8ca3aeedab9b1d11d1c713ce00
30070 F20101114_AACWUE cruzcasas_c_Page_104.pro
dab5516d79445f70f8cc1b7a0d560ec3
a5d6f55a774a61071c285e8415d6d90b8a36ccae
44655 F20101114_AACWTQ cruzcasas_c_Page_082.pro
4c87a3d94eb7fc139c84ddcb0cb0d8a4
200b3629762e5ca35eb4c7b50dc42ce8eca67b48
60685 F20101114_AACWUF cruzcasas_c_Page_106.pro
5d876613329a76823cb817eb094fdaab
f3818841d93fbd163fdf32ef2fb12b85c9c08201
17575 F20101114_AACWTR cruzcasas_c_Page_083.pro
351114fe33a304221f7bea4a46207cfc
c3717f005ab1e14eb3b413393b1202b48f5854a1
7023 F20101114_AACXAA cruzcasas_c_Page_091thm.jpg
2bda0a0d97a9e224c8e6dd76f6954ba4
c574545f8165492023bdb0c75fe16f2885e23507
465 F20101114_AACWUG cruzcasas_c_Page_001.txt
d3d9fd8cc5bb2d930c7561b69146648f
fd9648d591814c881bf0a87a7110c3d76877b8b3
52977 F20101114_AACWTS cruzcasas_c_Page_086.pro
514ffb449821ee7a0eda1eabe8024a9c
e0a6a6909fd4a468e68ca20412ca255cc89b6f2d
7529 F20101114_AACXAB cruzcasas_c_Page_014thm.jpg
f16482cb026d240088ec5e9ab8b8c329
e750ec95f26217991f39c3cfcdc1714e50089c5b
56785 F20101114_AACWTT cruzcasas_c_Page_088.pro
90d381459654d557d994fe9fdfd752aa
31c86665c1bc1799f4514e0dc0b3d6e9599a9acf
29725 F20101114_AACXAC cruzcasas_c_Page_010.QC.jpg
7bf03960dca9eea765e42feff4815336
832abf19851cca3636108fca15bcf6c7dcd905a7
104 F20101114_AACWUH cruzcasas_c_Page_002.txt
70b4ea14a1b85214ca8d55529a1635db
eec641051ed6790faec8debb99119746457b70db
11423 F20101114_AACWTU cruzcasas_c_Page_089.pro
459a01fd83f64f43121efad1671adbd2
0aef335e8c05f5ae13a37d33a8c9a03ae5ab2238
7695 F20101114_AACXAD cruzcasas_c_Page_082thm.jpg
e141f3f3c7e4de45e8d074c43b2e66dd
a61f1627a72406cac0d03465f379e0f5b4d87413
536 F20101114_AACWUI cruzcasas_c_Page_004.txt
db36b37a8e88cf51c6cd73e5330a681d
98877a72def6322be50a35344ca505f9a681dbd6
4211 F20101114_AACWTV cruzcasas_c_Page_090.pro
1847ea58674ab63ea58c214e54448d58
4df6a4dfc4be2980ca8e1d4b67e3978602e0da5e
31347 F20101114_AACXAE cruzcasas_c_Page_035.QC.jpg
78d8720d8a9758777d979a81067073dd
8e51d37f17822ad9b0090d08bc15df89fa1ce5f0
694 F20101114_AACWUJ cruzcasas_c_Page_006.txt
3d17fb945446ea6e143320c10db8d48f
953bd7981588b2079308fed41964ff340cab9ecd
7536 F20101114_AACWTW cruzcasas_c_Page_092.pro
c317cfb3b03354774f350edd29bee11f
a59577b37b299b07da5343a0510ca23e49fd0a79
8370 F20101114_AACXAF cruzcasas_c_Page_065thm.jpg
e4c9c6cb5dd8fea216d686668eb478cb
eb6fb376400f39642f7f9dc80696cc39b493be74
2309 F20101114_AACWUK cruzcasas_c_Page_007.txt
90de246fefc0fbd0c74f623ad5b8b2e2
1036e93a330099b3aeded910ba669300fc97ed24
1908 F20101114_AACWTX cruzcasas_c_Page_093.pro
5a040e0c18afe5e05dde7cec0afb36c3
71d82a2c8b4995336f8ffc58b8bb1490c58455a2
7918 F20101114_AACXAG cruzcasas_c_Page_018thm.jpg
54bf330e8b409575504ad85543af3c34
076961011cbc9e4f8394ea243b38b390483ebe03
1818 F20101114_AACWVA cruzcasas_c_Page_028.txt
8451ca42e7e57770266d7be6117eb853
869a34134055c35d602031448345bf71241a9255
2463 F20101114_AACWUL cruzcasas_c_Page_008.txt
2230561ff5da39ca2c2d18758b735b27
7536381eb00c0f725031fefdbfcfca78a515c12a
6557 F20101114_AACWTY cruzcasas_c_Page_095.pro
606f711912856b02133dbe661d62b89b
1d014dbc7c1d213d647730b70e760529e8d3497a
36429 F20101114_AACXAH cruzcasas_c_Page_008.QC.jpg
592884ce6931d482de6439b32ffabc3c
61ccfeaf350a72320b935f89919ae699e63fce8f
1981 F20101114_AACWVB cruzcasas_c_Page_029.txt
f58876db2637f2bb15f1de84baf4d1f3
e40f13fadf56d85a8fa12a44b978c24e5d1ca15c
1410 F20101114_AACWUM cruzcasas_c_Page_009.txt
a4272cdb40c7447362f708593f3f73a1
5a6a96381c86a7e93ecdc8a3f7b4a2c4796a044a
2205 F20101114_AACWTZ cruzcasas_c_Page_096.pro
6cd7679a8abb11e250694066151bc368
4ee8487dc31e159a79d0179e80e8fce77c8bce62
3259 F20101114_AACXAI cruzcasas_c_Page_083thm.jpg
432cd32c563f089e75f3059d3441d9b6
4e9d33de0f17a30d1113592234b8fa9c36dfd34b
2071 F20101114_AACWVC cruzcasas_c_Page_030.txt
26b86674ceb0f0808daa82535d998ca0
27a2fb62b92ce52922b2bada66042c9c32bcf931
1996 F20101114_AACWUN cruzcasas_c_Page_010.txt
555203861643e0b088abc4020f36407f
8ea4bcddda0a1ff1ab4a314bbd2ee1dcba4ea8c9
7887 F20101114_AACXAJ cruzcasas_c_Page_007thm.jpg
2557b97bbe6476027b97c78daa18ffef
25f08766231fa8fd269029bd1c6cae5c35463e2f
2360 F20101114_AACWVD cruzcasas_c_Page_032.txt
76e07e3d07f38f0fec410f045450f59b
a12716ad983916f487bc7faaae38a37b10476c26
1225 F20101114_AACWUO cruzcasas_c_Page_011.txt
5621bb28d81e328733c7e9969e936589
09c38f65b5fdd2a354163d87d68708eaa4ab3967
7796 F20101114_AACXAK cruzcasas_c_Page_030thm.jpg
ea566e6fb5babdf85006e0f8ae7588d3
33004c88e344541b4ec9568c582bb1d26ad68b02
2134 F20101114_AACWVE cruzcasas_c_Page_034.txt
91ddfbfc1c4f912c03c7c0699ae66d00
faf773a0a80dbf5a2a9a05ba8bd4c4584264db61
2177 F20101114_AACWUP cruzcasas_c_Page_012.txt
d291c3e3d25bfc900aa36ff15453a776
0c73d91ffc0d41bf3d8f1c7cfc1db35ad87fbbbd
36968 F20101114_AACXAL cruzcasas_c_Page_019.QC.jpg
e06d509d524976938b74055911c8c56b
b4c17646d8da2174891a154e823e1efa5337e11d
2284 F20101114_AACWVF cruzcasas_c_Page_036.txt
09f795be1bbe89685eaf17197ed061bc
e599bf7a36489af34cb987dda523f9b201cc7920
2186 F20101114_AACWUQ cruzcasas_c_Page_013.txt
d3822897b8f7ac4d7d124cf60d397c3b
cb6eecc5aa5218e823d70cdb9c489b1b48be3b43
8649 F20101114_AACXAM cruzcasas_c_Page_016thm.jpg
f98391cdc6df38b9b20397c0f05c5ef8
ba09222e4309c1a773a4f27d592ec08f13f26fd5
624 F20101114_AACWVG cruzcasas_c_Page_040.txt
8646aa6324a56ae143205e86f7c46903
47ce4fdc32286a77c04d0902562e6c63688d3561
256 F20101114_AACWUR cruzcasas_c_Page_015.txt
691c32f262f16090d8f7fc5d3ee3c92b
45ba271120a76b6f04468aa86b4f9992b385ad17
30691 F20101114_AACXBA cruzcasas_c_Page_023.QC.jpg
387843ca8b0862e7e29b426cc642bbd4
a4a8a3be7b33c64e7618595f9614373d2413f18e
24265 F20101114_AACXAN cruzcasas_c_Page_075.QC.jpg
c314eb800a956a5a4a2ae47cd743523e
c745a757a4f745ca156bf41cd49e02014f3e7e2f
163 F20101114_AACWVH cruzcasas_c_Page_041.txt
4e9ccc89de6617caddfd5316ab1e61c8
f749ab06b26661fb6593f005d810a08b4cde0cdc
2146 F20101114_AACWUS cruzcasas_c_Page_016.txt
6f12421a16506d5a780e4d801621849b
9ad81276aeeb8eb468cd5bbbdcaa4b5b353d54a3
31912 F20101114_AACXBB cruzcasas_c_Page_024.QC.jpg
38eacce05379ebdabfcc006259ab6468
c8f7dc5ae714ef2bb1aadfb064f365a60550b0f6
163340 F20101114_AACXAO UFE0021137_00001.xml FULL
8e3794a8c78c49665a59c783ec5c027f
0b97269451a4537ff1b6f1774ed7cb82d7ce5fba
1470 F20101114_AACWUT cruzcasas_c_Page_018.txt
45fb713b4660bfadae5cdafe4810aec8
c2172245d94052607b4b03d72b9a2e73b512bd8b
6353 F20101114_AACXBC cruzcasas_c_Page_025thm.jpg
6517bc2c7f7e93912b42914f0e29e779
674ea8312ea99df706428eaab3ff99bd5472b36f
8478 F20101114_AACXAP cruzcasas_c_Page_001.QC.jpg
5a0f2a35b162870fdcb551bcf81927fa
1342b98e00e6c5d79a8e7056d54b02cb5296bfff
213 F20101114_AACWVI cruzcasas_c_Page_042.txt
ee62d188b95bf8d9b6e03e487541ac68
719891d9f6da394929aa6632d10e6474949b3e90
2203 F20101114_AACWUU cruzcasas_c_Page_019.txt
9469d9cfa44c9d26360ced474b66c43b
f8202f02c5aa6230648fe5c1551fd42805d14edd
23901 F20101114_AACXBD cruzcasas_c_Page_025.QC.jpg
3c14b19f94a8f7b1334ca908981624eb
2ce0a8dfc210d918bdcdbe2ee63081b8f3fb3ddd
12169 F20101114_AACXAQ cruzcasas_c_Page_006.QC.jpg
c8aa0b1ebbc18cda40122ecc7bc6284b
6196a5fb3a6a450b18e4e0a5eb9379e3a3ec897a
651 F20101114_AACWVJ cruzcasas_c_Page_043.txt
a68b70ce07ea1d1ecad18b27364f5006
58dbb10153d83a1b26eb3f5c21778c139d754c14
2152 F20101114_AACWUV cruzcasas_c_Page_020.txt
1325c9b330afd16820c28014108a8dae
c87c3bee2ff7bb1a951b062064ce2638179bc00e
7486 F20101114_AACXBE cruzcasas_c_Page_027thm.jpg
c34bc3bb98643a3f274b1c3252339aab
e4b21d74a131f3669aa7d88bf81c9eaa767a1d0a
8585 F20101114_AACXAR cruzcasas_c_Page_008thm.jpg
a58277aa5a5bf857cb61c0a1c1e46b85
b58a814200b54d4ba8c0d0a59dd9a695415e63d6
484 F20101114_AACWVK cruzcasas_c_Page_045.txt
32c900278fb0db2164f16357831aeae6
303678dcb8b2eae7264b56f96ca5a62dbfef8e75
F20101114_AACWUW cruzcasas_c_Page_021.txt
cb3294b6407dc24b3d8fc50f9f1fde14
c2a76d38cfa8dd292cbb0f864e524066d00e49c1
30461 F20101114_AACXBF cruzcasas_c_Page_027.QC.jpg
e5b7166bd5b8e527e7a62ff67d7ad8dd
0bf8acbdb168289451139fa2e59f869772e06c96
1701 F20101114_AACWWA cruzcasas_c_Page_073.txt
771f1b1ad872cc22022f5bf6a1b09fdb
4f413370ad31307b402dc95304a03e7436935b16
1063 F20101114_AACWVL cruzcasas_c_Page_046.txt
fe309fa5cb2a64c5c380ec177778dc7c
81f9d4639d8d5649e30cd2f9472734dd19fef36e
1917 F20101114_AACWUX cruzcasas_c_Page_023.txt
41156d3f9b0cc54882544c18a52bf456
7f6db4d81ed7d179bebe042c45edac3e1fc15edf
27333 F20101114_AACXBG cruzcasas_c_Page_038.QC.jpg
1d681c56c930d4a9a81c3046fc602778
cc941b46570ca6ba7d491a94915bbb1f6a5ede1e
1129 F20101114_AACWWB cruzcasas_c_Page_074.txt
9ab0f72392a7064c79adc1bbc8212201
6939233de671e06dbf0ef75d6dc810b318583410
8979 F20101114_AACXAS cruzcasas_c_Page_013thm.jpg
81a0e16d7bacdf27fc56fa116012fed6
0815c5f3b76c167645eb61870b563e8601d522d1
1934 F20101114_AACWVM cruzcasas_c_Page_050.txt
e68a48c55559ba94147827eb6814a10c
9db9c46853f1624bf838eff4aff043cb781d102b
1517 F20101114_AACWUY cruzcasas_c_Page_025.txt
809327a4debfb8553a26c4829f2e02ef
be851d23a25fed6031f397ce5258afc615962652
36113 F20101114_AACXBH cruzcasas_c_Page_040.QC.jpg
516f2370344072b2bb4694762dbeb533
538ce005d0858ab11a9cc08c9dc78faddb1a5645
2577 F20101114_AACWWC cruzcasas_c_Page_075.txt
96eaa9686ad65460c7d53d58b1b9e473
f0ea4084c8adfe4159c3d6b3e158fbff04d9b19b
35620 F20101114_AACXAT cruzcasas_c_Page_013.QC.jpg
0976bd5b338976bf0782f230752bf71d
453b7d1c07f2d29b09b45560f436e5a66d8b7da0
2059 F20101114_AACWVN cruzcasas_c_Page_051.txt
428fd27ec5301f8174fbb42bcb0f415b
631bb15628e0d0a1d955d3f554c4ae05eed202d3
2001 F20101114_AACWUZ cruzcasas_c_Page_026.txt
02b069a93825b2596e2cbd02899d5de4
808c97c0fa62634d1cf720de894e96d0d362b45f
8045 F20101114_AACXBI cruzcasas_c_Page_041thm.jpg
311001140ab444a9a5242885efbf7a31
aa977702102bf739343ecd4b65e07276e7b091c7
7146 F20101114_AACXAU cruzcasas_c_Page_017thm.jpg
72c3348cc5b6f369cecebe4cc0ef0188
6c5839e2d426bde92fa21f0a3400eeebcef7593c
2099 F20101114_AACWVO cruzcasas_c_Page_052.txt
1826908bd91f1d398567f16729f9d43e
ffce206ff3182bd3caa7417c9e0618640860803d
25556 F20101114_AACXBJ cruzcasas_c_Page_042.QC.jpg
4aa23681d82365c2af1d8dc9ef77ae18
0eae895f47455dd81b0b5e8b3c6e895f03eef84e
2013 F20101114_AACWWD cruzcasas_c_Page_076.txt
d97c48e842191b45d7c6eba540824bb6
d6943500966b14a281e0896423befc552c5237fb
30661 F20101114_AACXAV cruzcasas_c_Page_018.QC.jpg
eaa1492892574c435d184e7383a74893
e3c50d10d1e7dea1c344807025c0613b15f2fccd
2173 F20101114_AACWVP cruzcasas_c_Page_054.txt
ca64082a1e1cd6237a98433c78124f0f
837dbfb91c206508cc5945c929bea5e46f19b743
7850 F20101114_AACXBK cruzcasas_c_Page_044thm.jpg
b5a7b427e3023c5e025184ee4725cc54
4018d7926d951fc2216a2f7128987e359bfe38ed
1863 F20101114_AACWWE cruzcasas_c_Page_077.txt
719615310e54bce4dc45d6d9f532957c
23d72a0bde67e12b6ea8870bf5849784d74aec09
8645 F20101114_AACXAW cruzcasas_c_Page_020thm.jpg
9f6798686f1ed315dab5cfbff9e4a0f6
d4255629194f30504edc24bfb2cf82ea3faea098
631 F20101114_AACWVQ cruzcasas_c_Page_055.txt
eda0c36b078d51307f23115198e7b170
9d5351cb0d7eb54ee2df3fea8a6271e324a81a89
31762 F20101114_AACXBL cruzcasas_c_Page_044.QC.jpg
1b4068ad743f7c7b7dfc2171ae82df42
7232e2f6439553bdc8997771f82ed5ec7cd22b73
1411 F20101114_AACWWF cruzcasas_c_Page_078.txt
2716e140ebc738e492c96db68a6a82c4
cc2958bde03e85fc9a83ed012f2cc7a134953da5
35257 F20101114_AACXAX cruzcasas_c_Page_020.QC.jpg
54419f1343bb5272074baf5899c423fd
f236f073faf091ada2bcb8e0ff957ff838099e94
F20101114_AACWVR cruzcasas_c_Page_056.txt
c5fad5856c12888c87c4c7284cacd5b3
b22f445a75b122063da0f8e6c0df27a948664a20
9193 F20101114_AACXCA cruzcasas_c_Page_069thm.jpg
28583d0240bdc45730cf0702b33cf086
8c54c5d9b46a49b8a6cc3f95b3f8d03d68d84da8
22197 F20101114_AACXBM cruzcasas_c_Page_045.QC.jpg
5c1cf4f587d9f99b2001bb1bd2078f01
d7f8bcbc7d43beff4820b06b259e5ebb8675ab00
1699 F20101114_AACWWG cruzcasas_c_Page_079.txt
debe8578fd80ae2b29b05f26c1f01f08
9b4a516fe2f61fc7937b6c3e13d812bcdea80846
9048 F20101114_AACXAY cruzcasas_c_Page_022thm.jpg
e48f6cee65116539c4c090a65598fec1
30621317b8938eedb8f6779bad87d8aa73b3ed47
1841 F20101114_AACWVS cruzcasas_c_Page_058.txt
c240f2c1ea4af006d4557018d9e2b65c
5d30a5127b9b9f6574c8a168bf6e44b3c2d62896
7273 F20101114_AACXCB cruzcasas_c_Page_071thm.jpg
4b4c8817ae9f4a13740ce7165ef32573
162b42399b905945d64938fe231b7ebfcae617e1
29325 F20101114_AACXBN cruzcasas_c_Page_046.QC.jpg
f2a448905f84409c40725094a33c49ba
d43d9627d89a82c3f84b29daaa9e79e4e44a4805
1864 F20101114_AACWWH cruzcasas_c_Page_081.txt
b076ddae837d177af2353cfce72370bc
6114f81bc36882f6c506724f60e18cfedd576ac2
8142 F20101114_AACXAZ cruzcasas_c_Page_023thm.jpg
3685e7cb7eff9b00929259e013a666a0
5f605abec2d94d37db670dea7ef6c913353f4440
1995 F20101114_AACWVT cruzcasas_c_Page_062.txt
d21f10a78b65bd067449c008645e8e24
2c95cf636ca319c95f0a9f7a0ead2997311e0c6e
7258 F20101114_AACXCC cruzcasas_c_Page_072thm.jpg
c863f6ef774f2ba1a7574e17af332f3a
7800f55ff99786e4f5cbb82ba1d3a289dd8c4f50
9613 F20101114_AACXBO cruzcasas_c_Page_047thm.jpg
02a293ed85d61c950c713a11b8e5d80f
44c3d5c4ae0a8e5abc84f0ddbb389d8a6239ad6f
1848 F20101114_AACWWI cruzcasas_c_Page_082.txt
e86bc4174f76c78a1fbf65ce0677257f
6d649b36a95060415d796fdf2e38edbd16057e26
2193 F20101114_AACWVU cruzcasas_c_Page_065.txt
c95a76e8c4bab2c12e674c11422534a4
c13b56d0183e9e227a16e1187e26a5be5b504087
26035 F20101114_AACXCD cruzcasas_c_Page_072.QC.jpg
eae476df0106e9b56a16271759d67c11
fea114b248b72acd382b349644a322cfa84f28e3
8362 F20101114_AACXBP cruzcasas_c_Page_049thm.jpg
c4d8452bdb32f6277a0c1210e76adc62
280919a2fcef8042641a20141d10fd2cad7b06d6
2014 F20101114_AACWVV cruzcasas_c_Page_068.txt
75b563e916c1abafa6f2efd46810252e
7b8e65a91c2f7e4e3c666a3b8fd9ef4f797e575f
22037 F20101114_AACXCE cruzcasas_c_Page_073.QC.jpg
cd5e1214e1fb417e863ceac5c7a2e2c4
0d1a15e60c15214ae61b2244835319fb21804b37
33860 F20101114_AACXBQ cruzcasas_c_Page_049.QC.jpg
d572a934a072c1b50300e1fa77a2fe72
3bf183364c1c0fbec357c77733b8932dfdba7a7b
2260 F20101114_AACWWJ cruzcasas_c_Page_084.txt
abcd5e9965f50af8167fdf5011ab6bc8
57e7b92b7e8fd7aaabae70e4020c35179a329468
2191 F20101114_AACWVW cruzcasas_c_Page_069.txt
dbba95b68b9b6343e85552f7f5f5619f
0d49b65394d8b5afcab998ef9b2db674dd8f3add
28469 F20101114_AACXCF cruzcasas_c_Page_081.QC.jpg
68b314a23ec80016a33ae078ee1f360a
1fa647659105d0de9d6fa063a66ea30deb06cd18
32501 F20101114_AACXBR cruzcasas_c_Page_056.QC.jpg
d149c34e91ba7c10c21b3405b4e5846f
99f8e516f3a3ee7f8554864d634e3f6d1c3a9d6c
F20101114_AACWWK cruzcasas_c_Page_085.txt
aeeca3b0cd732b309d8b36bbe8f1efa8
c60bf4d9bb10f806299df5d03fae713320b9246d
2199 F20101114_AACWVX cruzcasas_c_Page_070.txt
201838850392644f2fef99dd6d70f9e6
575f4bd7131945f4026fa0380b0b158a2a4dc25c
9100 F20101114_AACXCG cruzcasas_c_Page_084thm.jpg
e534d02c11f0ded24e810bb4394fae4a
1ea603fb9a8016002a72b96561d0b4213d150dd5
5322201 F20101114_AACWXA cruzcasas_c.pdf
5246ab1ca42c07937416ab0106da9956
ebd18eaddea506f53c66ed064051474290f4fa48
29424 F20101114_AACXBS cruzcasas_c_Page_057.QC.jpg
f11bedda688bfa1ea9d783cbaa9ca03b
20d4619746c2527b984e2814d158cf3bdc4b8056
2226 F20101114_AACWWL cruzcasas_c_Page_088.txt
2d4a90f7ab20d14a65662f9175f96ae9
d8de2e950e95e0e4f838debd4a4a80ddd24c8beb
2332 F20101114_AACWVY cruzcasas_c_Page_071.txt
19311c641d15013b3696d0ccef1627f5
804d6f06344b6c0b0e1567cc0459126becc5c63c
35861 F20101114_AACXCH cruzcasas_c_Page_085.QC.jpg
58d660196e67e0e4e2f7f87511b7ca36
9de5422cb36501c2376bc0fd3ce77cffcce7e22a
33433 F20101114_AACWXB cruzcasas_c_Page_034.QC.jpg
8a0f1ab7273119c3f1748e51bddbd313
adb8c67e2eb68903b231e60d0d3f81b9b57ae759
463 F20101114_AACWWM cruzcasas_c_Page_089.txt
fbad7493676f99bb4fc27fe8d7157bee
cad616a655da8034cfa0f92acaeeec8f5b1ad6af
1571 F20101114_AACWVZ cruzcasas_c_Page_072.txt
068ab46f399660a4bac95faa067a79ac
9b2da3bd7def9a88e1879a3de47ed397553c83ee
8922 F20101114_AACXCI cruzcasas_c_Page_087thm.jpg
212701a205c3c84e003a73e6f891a82b
71ebe90f99720419be27c0187999d9ab6570a811
8666 F20101114_AACWXC cruzcasas_c_Page_032thm.jpg
ccd494437122067813978e5e183b1ad7
6167ef331231861d5fdece127ea22138450ea303
7382 F20101114_AACXBT cruzcasas_c_Page_058thm.jpg
b2a40125afdb28775deadb02b44ac3f2
4abbe0232fb057d06adc217bb10e003fa466ad3c
190 F20101114_AACWWN cruzcasas_c_Page_090.txt
c1005f226c382f04bd9ecf22aab6ede6
3f065ddbb69cbbdf81655992b34ef40d64a889b2
8964 F20101114_AACXCJ cruzcasas_c_Page_088thm.jpg
688ed22e5499400730b030957bcdd173
47c89125f2932f9cb0a20ae0d727b58a1f98b3d1
7927 F20101114_AACWXD cruzcasas_c_Page_024thm.jpg
6e313f0dc6d6dc0271198eda28238419
2adb6ad308c2ff82bce41cc1f35956e09a4785d7
6267 F20101114_AACXBU cruzcasas_c_Page_059thm.jpg
b6b093b68115dcb1f3e7bb04aeb7af5f
4484f5c3af5a250cc32651f5da7a4a58da27760a
616 F20101114_AACWWO cruzcasas_c_Page_091.txt
3b985cade812e09ac4107518403fddd3
3296fa7ad36867ab043b2f3b0a38e33018b90eb5
15071 F20101114_AACXCK cruzcasas_c_Page_090.QC.jpg
38a847eb2c16f84f72dc01a197a2e98c
1371b23b2e33e49df93a553668077664264dffbe
6950 F20101114_AACWXE cruzcasas_c_Page_074thm.jpg
391423beee52f86833d29bf93d9b489c
25242a601d7fd39052ce3acbebee014d3754cc8a
23318 F20101114_AACXBV cruzcasas_c_Page_059.QC.jpg
c37f90048112a29dbf4fa8838674bd53
588f45b653f5b4762213c6df1c838b89202e54f7
364 F20101114_AACWWP cruzcasas_c_Page_092.txt
d5921fd22a864de4c7eecdc9bae02a80
1643fd1a68610746ca6b08211a164c53bb996ada
7383 F20101114_AACXCL cruzcasas_c_Page_092thm.jpg
b6bb659c836ded46ee0128afc0ffb5f9
b49822b9fd19570bf894f380b81a108e9c56dcab
12176 F20101114_AACWXF cruzcasas_c_Page_055.QC.jpg
7ed5bbf0598c60ca7ab73fb56c7de313
88f4d34a5a141f8357b912cdde2d17eb07957bb6
22184 F20101114_AACXBW cruzcasas_c_Page_061.QC.jpg
09559b9d2c722be50d74bd8a3086c2d2
943a6b187913e11a369710ebc072f0775e81498c
445 F20101114_AACWWQ cruzcasas_c_Page_095.txt
333e9095656080576df00b10def2934c
26c04d95da94db6baa40c9c254f483cf70c9ad3e
21655 F20101114_AACXCM cruzcasas_c_Page_092.QC.jpg
b2e4f8c1e42e0d3fe88dafb08b72e7b9
317be62dbaae1c000b2d9a64bfc2bd4a836d5435
8780 F20101114_AACWXG cruzcasas_c_Page_053thm.jpg
36d2d9559ab581a6bd0486a22c945f53
33df399441471570fc071b03b6b90aea5ef3b892
36342 F20101114_AACXBX cruzcasas_c_Page_064.QC.jpg
2908102dc04faac3a79c479a71fcf606
b5b08b85d1327502679366fb2a0d4272f83f67b1
191 F20101114_AACWWR cruzcasas_c_Page_097.txt
e8cdf4d65c1c91a2844a2b3e0f46806f
0efe607b69a127b3e19a5d68940473535915ffc1
7988 F20101114_AACXCN cruzcasas_c_Page_094thm.jpg
2553d48b65933ae3f8e4cf761dbd7d05
3850903790c5f5f9b6203af844b92325e8789163
29724 F20101114_AACWXH cruzcasas_c_Page_063.QC.jpg
06c4304bb2ae9a7024e298fef1fceb98
a4b6d753c9d4c23bef6ad126269f50319d2b4cfb
2888 F20101114_AACXBY cruzcasas_c_Page_066thm.jpg
98d907d568467096176a89142a73fa62
6ef5d39dcb30071b8a41d84ac229aa5f41551623
1825 F20101114_AACWWS cruzcasas_c_Page_098.txt
16cc65d436dc0a60ca09bbe589b9b6b0
f72a48165d6a53cce45d76e0e0ff616922339fdf
8943 F20101114_AACXCO cruzcasas_c_Page_095thm.jpg
48f4a66f22494c3342b89d9d4d727f4a
2264d5eee3daa80c2227ecb57a115b39c601d343
7113 F20101114_AACWXI cruzcasas_c_Page_102thm.jpg
5edcda1a5b8f2608884a1ca19d99b649
ccd5077fa303dbde7483bd0710dfdb83ba6c4476
7791 F20101114_AACXBZ cruzcasas_c_Page_067thm.jpg
c08d15ed54994eb20b272eb2b693dd9b
9c8ba1c1705f922937ffc95f63c9e379feab7a20
1694 F20101114_AACWWT cruzcasas_c_Page_099.txt
24b15f739ff41863fa5dc35532dc1776
5057d83bac9dd897b8dc41f86b528783c346c5c2
1010740 F20101114_AACWAA cruzcasas_c_Page_038.jp2
404d7619ef5952c9f925b51476d2b9b0
217c11d366a27fdd5be01f73e09656baef0d70d1
6527 F20101114_AACXCP cruzcasas_c_Page_096.QC.jpg
00b2d5ff8624c2574a29ea47cb4fb542
c50d380fb5c62eb11bdbb7f670dcfd38697d86f7
30864 F20101114_AACWXJ cruzcasas_c_Page_103.QC.jpg
15be16c361eaae31f0109e7b6ba17d56
90e280ce4b72d7885b25c3fffc708d27a94d0c25
1680 F20101114_AACWWU cruzcasas_c_Page_100.txt
d1a572c3d1b9c22776d267c510420f69
8be7fa59e370e07c1c7d78e2dc20e960180a85ef
F20101114_AACWAB cruzcasas_c_Page_012.tif
4a0b7c18f2c8fc613e085b57accfa88e
19803a5333d7efa55cfeb0f2176014f54aef15f0
3939 F20101114_AACXCQ cruzcasas_c_Page_098thm.jpg
46a6d2d3559597c21366817261da9aa5
1e35d86bd7f7d96a4a3afa4d030823514140b7ee
1405 F20101114_AACWWV cruzcasas_c_Page_101.txt
22d40d64be4f8af162f19116da8d2f48
666605c7e62f8f9e2a48412792ee5a13c903568e
1051922 F20101114_AACWAC cruzcasas_c_Page_042.jp2
799c6462673006f20b25ba46fc5bcb12
c284f88da0c12a6dcc06954e5bc0de083442a1c2
17171 F20101114_AACXCR cruzcasas_c_Page_099.QC.jpg
6697e7e8fe1fcebbd3cd5653a66d7a08
849841ba182bc1a474978261a3536af8ad590dbc
7480 F20101114_AACWXK cruzcasas_c_Page_048thm.jpg
b75ab701a7f8cb8d8b77ca3bb96b01d8
25cf40450318b300c1bf0eb7f2ca31ee124e0446
2183 F20101114_AACWWW cruzcasas_c_Page_102.txt
25f5748e17fdc2ebcca8c000aab96ed9
e4c3971f452f6161f115dcf6b95d95b0fd0acf73
522186 F20101114_AACWAD cruzcasas_c_Page_097.jp2
6699217e654b04ff208a45f6c67f12cc
f7c04f8334b427f3ebce14cdef0acf855668d170
20306 F20101114_AACWYA cruzcasas_c_Page_097.QC.jpg
6d7b4f5a3630147e0d923f928ae4affa
3e8964411dd0609cc11be4aab0e77e1e2d6c48bf
12844 F20101114_AACXCS cruzcasas_c_Page_100.QC.jpg
87a69a7c278e566d03d84940b535b602
c988b1e7aa247048264f1c03194a40af669e0151
6623 F20101114_AACWXL cruzcasas_c_Page_078thm.jpg
a59d0013c9a4e8db0634af9b87da1f9a
4076c4c1e7fed402616b935977fbd6c04833ff3b
1351 F20101114_AACWWX cruzcasas_c_Page_104.txt
a0e501e55de7b1452974662f5e22fbfd
dce68d93beb81731531ce27839985784a6ea8b35
8700 F20101114_AACWAE cruzcasas_c_Page_034thm.jpg
170b65f50e2b13bb963cd1bd28a57069
e38168e1690650e734ad7b55cc1794f14a7083a3
9460 F20101114_AACWYB cruzcasas_c_Page_068thm.jpg
31e1f9ae361f86aee5d6c6e6a97831cb
1d5789da9adde426917ace3e161f35afed794a1e
27945 F20101114_AACWXM cruzcasas_c_Page_026.QC.jpg
65adf5bd418f7fa418d618ceab9ab0ba
067bd32392e23067b2cc98572139d634163c9c4a
2532 F20101114_AACWWY cruzcasas_c_Page_105.txt
240463f955d3b675bb5096a377ae7388
2ecc6cf81e4ee1e2636590c83d040164e8b99cad
27674 F20101114_AACWAF cruzcasas_c_Page_048.QC.jpg
0eac01c04df941d30c0fc509b5e8b8a8
795a0b1d6e63c54358b1ef2eed2ba4b99a967d50
1526 F20101114_AACWYC cruzcasas_c_Page_015thm.jpg
d02f48423009bc92d8570a25c8958f78
d6bf5b7657f5293707bed373f6ddf40358ef6c02
6516 F20101114_AACWXN cruzcasas_c_Page_107.QC.jpg
1c66e4943f10757cb360639239b99a53
8334c93c1f2a8e4f573bbcea61f9f886f3b8d225
363 F20101114_AACWWZ cruzcasas_c_Page_107.txt
d5717026c457546303930e9d4d79e1d0
c156895db65bbd92f21805a4ceee9191d21d81e9
F20101114_AACWAG cruzcasas_c_Page_049.tif
df3599ff271c06aadcddbbac6c3cf0c1
a21b6c76376938fee30ba00b4922561fabc77251
7090 F20101114_AACWYD cruzcasas_c_Page_026thm.jpg
9e2f163b77bacd4296e547ad93d71054
30160990796c031a5775118de968b0674d7f88af
29984 F20101114_AACWXO cruzcasas_c_Page_077.QC.jpg
e270dc10474e4b46d055a4c25176ca2a
0df8a5c85ac2cef0f6e8b7f8442080c867ffa664
90336 F20101114_AACWAH cruzcasas_c_Page_077.jpg
e0cb8a0d87d4fca0926543ec890e2e71
2b9d4bb8e1cf0ad2ba167f40c522c40208f9d027
8939 F20101114_AACWYE cruzcasas_c_Page_036thm.jpg
0dd54717363b110d02182b7e78675e9f
0cd0786f214777fb2aeec76d3cac9f06365028fb
2928 F20101114_AACWXP cruzcasas_c_Page_055thm.jpg
6efa4cb49bb86a05ec22fb8ca23db12b
9e2183189526dbaf4f37b6d2cc069afc7cf9fa57
7759 F20101114_AACWAI cruzcasas_c_Page_060thm.jpg
6809840e16b61e93cec0447ae1914e91
15cf2e24ccf3bac7fe5e5bee4c9cab9cf42085b1
36165 F20101114_AACWYF cruzcasas_c_Page_069.QC.jpg
7a129dcb0ac546c0ec6c89bebc23c75f
257daeadaf4d973f9d9845f1cda3bab8efc998da
18006 F20101114_AACWXQ cruzcasas_c_Page_104.QC.jpg
4c979d0ed5cdcaa8c6f74324ceb7d178
9868a2b96950882a753ac4f47fa93e67c2349ad0
98008 F20101114_AACWAJ cruzcasas_c_Page_043.jpg
c454d7939e6ef618cd13eefdb63e14e9
1caab643a2bdbb39100214889af9992da43bd354
3698 F20101114_AACWYG cruzcasas_c_Page_100thm.jpg
d79885cc5a5d0a7d9ac20e850078e185
c249989f72cbc8e92a990e483594d1925d8d5928
5208 F20101114_AACWXR cruzcasas_c_Page_011thm.jpg
4b52260165356aa164ea919b84132e04
8f8e6fa70174cce6d62f6ff2acd1eee2c8933a05
54368 F20101114_AACWAK cruzcasas_c_Page_099.jpg
5b7db707e06106dbf2a485a356af03a1
13c130ff866f2e78e7a08312ec87f498087f5709
35721 F20101114_AACWYH cruzcasas_c_Page_022.QC.jpg
bf79a51deb99c591cb430121b5c41ebf
04b181616851e0407a8ec3006575ad6608f0e0d1
22390 F20101114_AACWXS cruzcasas_c_Page_028.QC.jpg
fd2c3a1224bfd1c39e64a692fc32856c
1d3be282fee80d447dc45affdc63e182b61ae9a7
1418 F20101114_AACWAL cruzcasas_c_Page_059.txt
02c8dae1253ed7ca214cea56e7a65935
bd1b2eb4adcc106e9c488d04e802038fbb9467ab
4222 F20101114_AACWYI cruzcasas_c_Page_099thm.jpg
789beb85a2a82a4dacf40a4295ea0e2e
446eb897e8bea6efef3421abc0d46bc78d3ff877
1244 F20101114_AACWXT cruzcasas_c_Page_002.QC.jpg
ab51fcd4b9916e528fd238434b73f83f
0b11c35ff99f79e19b01ccdbc124f99496d3cd3f
1051944 F20101114_AACWBA cruzcasas_c_Page_105.jp2
16cb135cec069fb11caf02facd38f541
eb7db795fc02ad86c54c7a361932f9f16eed6a6b
43735 F20101114_AACWAM cruzcasas_c_Page_058.pro
2023239fa91c168d2e89a8825f9a81ab
96fef9f3a8a04109607eedd6126ffb9bc2201b15
8846 F20101114_AACWYJ cruzcasas_c_Page_106thm.jpg
686eb2d2253254f10300e79b11949806
18667e92386ff395b2c66e33f3b9c82fc2112bcb
34954 F20101114_AACWXU cruzcasas_c_Page_012.QC.jpg
a335e47a126dfdecb98753e2474a7e4e
3a3f6793d4746746de1a9f6b901732615204a9c6
9150 F20101114_AACWBB cruzcasas_c_Page_019thm.jpg
d60ea719393c5bcb7d31ee8d858d4ebe
35c15a5970be6df4a43bb0db7526cb324e18ab73
F20101114_AACWAN cruzcasas_c_Page_059.tif
eb2f12912416f3a1eaf3a92be832e125
265248ef006ecdf8aa01dc0b5f93454f917d54aa
8887 F20101114_AACWYK cruzcasas_c_Page_086thm.jpg
99ba17eddd6e98d83be7f3464cc57774
5f3a2288e85e814a0155cdf7365d97f38df2b839
29610 F20101114_AACWXV cruzcasas_c_Page_062.QC.jpg
1dbf356ba824ef5bb05ab9a690d10a29
1441ff95d35865cd02648d2facdaaec8beb53134
F20101114_AACWBC cruzcasas_c_Page_107.tif
342380a74009c0e4b88cc301ea32b8f2
09a3b813b61e2dd5fc61d1700edd3db5852d110f
205630 F20101114_AACWAO cruzcasas_c_Page_096.jp2
78165d5fb5c6b3a38927a49abdcf959c
8aa76775e8a78ea6dd93bec3db9a8853a7708559
22837 F20101114_AACWXW cruzcasas_c_Page_094.QC.jpg
0e094bc5f5902843a194a405e2ded9a8
2d36966c075c61e2cfe8e05c532b0f579ae14b71
7697 F20101114_AACWBD cruzcasas_c_Page_063thm.jpg
f9bb21696830835493e9a90bde12cea2
cf0ea0b70fb59075618a31b521e4c958ec834895
35524 F20101114_AACWAP cruzcasas_c_Page_065.QC.jpg
26736e653941edda52758097c6e10c7f
00f0baf95eaf5825979b2fd73c353219fc7e2e11
1403 F20101114_AACWYL cruzcasas_c_Page_033thm.jpg
39c27031f278dc154482c9cb3756389f
2fb58b16f726d30959948c0248b604ef5d0e13ab
26274 F20101114_AACWXX cruzcasas_c_Page_005.QC.jpg
ff95668215df0158fe49065554ee3d42
fb01d7ba311575d95d62158d7b979178830bf0c3
2157 F20101114_AACWBE cruzcasas_c_Page_053.txt
69e50dec7eeca2c223c4583a064a7036
8e0882ef69aaa2cd9178ef72cc3bb4d2f51ab6d5
7302 F20101114_AACWAQ cruzcasas_c_Page_079thm.jpg
3e1a7f6da21429ebc9c4d25a08c62058
22193529378a77c90ad4b3be1ea7c9df9a0b77c4
7326 F20101114_AACWZA cruzcasas_c_Page_103thm.jpg
5acdb46cfe70957fbbcdd7c5967ec495
d827e938922ad995ee83d480f1929d29d0afadfa
35334 F20101114_AACWYM cruzcasas_c_Page_106.QC.jpg
2e0aeb1bc1ae652c8e0bafcafbb9969f
21497c0df063cf8f5327942bbe33697cc6702a33
31422 F20101114_AACWXY cruzcasas_c_Page_043.QC.jpg
d4b0ca282769e3ccfe403f29644cc0d4
251d08600c6e530c8c9352531055e556e22b9e11
27584 F20101114_AACWBF cruzcasas_c_Page_089.jp2
090b83b1616aae0846f321e972bd2246
97ef19de1570e128ccfa8ded4c4d3dafc88fb41b
F20101114_AACWAR cruzcasas_c_Page_105.tif
627de1aaf4a45afaa38a284a30ce14ca
6f9ac9a634cd438e04fa88d6a281f04a76cdbd64
11504 F20101114_AACWZB cruzcasas_c_Page_066.QC.jpg
f25bae56666dbc2ba53f3eeb2a91e947
86f581d764edb26db3d01b335f45115cfbbce4d2
35798 F20101114_AACWYN cruzcasas_c_Page_054.QC.jpg
8855cbbac3fdd9319c684e9a5dc9b3a4
839a185d12c2546847d554186e4ad22d31549b9c
32779 F20101114_AACWXZ cruzcasas_c_Page_047.QC.jpg
1b4d243a8283d023e13e79ac3abfccd6
8b48a9980a75a61a3ca2a29e90d62b49a366af44
968914 F20101114_AACWBG cruzcasas_c_Page_095.jp2
599854069be16ab10e289665058f285b
6ccb0fa2d648514fb8cbb90bc0652c6c9266ded2
2426 F20101114_AACWAS cruzcasas_c_Page_106.txt
ee5cfe5788690a2968fee8e5c4969710
b087e8b4df0a009db908268393fb9d00d6ba984c
35567 F20101114_AACWZC cruzcasas_c_Page_105.QC.jpg
65da76092300649719aab19ff03db20b
0a833d66d902eb1085d1e106b11932e935e26e53
5989 F20101114_AACWYO cruzcasas_c_Page_028thm.jpg
95f0ba7e0ef1241e475f8c0c667a67e1
5fa26d9b754ed9ae682974216f10ef8ecb2b3cab
8565 F20101114_AACWBH cruzcasas_c_Page_043thm.jpg
bb795f5e7041ae1538dddfda18ee074f
7144ae6d897f9f4b4fae6b57feb977ce69ca7fc1
8199 F20101114_AACWZD cruzcasas_c_Page_038thm.jpg
1f05dc17df183e917e1475774ff54609
babfcbe6a2c9885d98379ac00f990d3490068edf
36973 F20101114_AACWYP cruzcasas_c_Page_036.QC.jpg
b6bb0165822925f4e963844ad6d029ad
2a9666ffe687a1c0ee65c4e8ab9b0d89cf9ab63b
1743 F20101114_AACWBI cruzcasas_c_Page_017.txt
d5d8f6d46ef0b9b638a6f61b3e7d00e5
a4b7b0aac960ea633835243685f65299a6df0197
6247 F20101114_AACWAT cruzcasas_c_Page_097thm.jpg
02ee5d4f2db1d2322a797cd6afb0754b
b158471cde1a720eb970f8579083359539abefba
9928 F20101114_AACWZE cruzcasas_c_Page_004.QC.jpg
4c47c50fdaaa7917a06ee4e83df116ed
5c0c85659c9797583094b27a0b276fa9f3a4a782
6113 F20101114_AACWYQ cruzcasas_c_Page_073thm.jpg
9bd7293a74c7ce1c6c54ba727b4c6635
4faabbedd1ada13fa2b2c1fb328556f93bce4f46
109 F20101114_AACWBJ cruzcasas_c_Page_003.txt
018d074496a01472eb9403899cc2077d
3fe54a9d6955e591d5b297304bb85d96731d898d
2022 F20101114_AACWAU cruzcasas_c_Page_014.txt
a35dc79b5ab5dc0ab7168ffbc599155d
ce8fc3d0990ad8b96c0c0e566e1a384ddf3189a3
26080 F20101114_AACWZF cruzcasas_c_Page_091.QC.jpg
f9779d122d4d371bf047f83cbdf50a16
14c6c7a7db5c3e84f621e5ad84aaef280ac00df8
28580 F20101114_AACWYR cruzcasas_c_Page_017.QC.jpg
de4a0651368358b297cd3b3efbdcee2e
ff374160d649456d94710f0b292dbbd5dbdb58ac
33341 F20101114_AACWBK cruzcasas_c_Page_050.QC.jpg
1de98bcf92f36c857366ff44713ace59
bc86fc88c64be29015b805e77b59defa3077f1b3
109570 F20101114_AACWAV cruzcasas_c_Page_021.jpg
41357d8a15987311dcbb4b19c86d91c6
1c63e54c6e60d2c37fd15ab187e983e9aa71861e
7607 F20101114_AACWZG cruzcasas_c_Page_010thm.jpg
1170b54834717ae7ea5f26c49841493c
c979b41b9f38de6ac7d3852aee3aab37745c229c
29689 F20101114_AACWYS cruzcasas_c_Page_071.QC.jpg
0a8bdda0ba9e71c3f6d62afdcd9bda75
af79d983b505bb35181823f23c34f88ff583a7d4
45522 F20101114_AACWBL cruzcasas_c_Page_051.pro
473bd662fcd235fd8d413ec446d2a905
7608ec921cafa6994c57a19f34d34c3d772e9be3
7765 F20101114_AACWAW cruzcasas_c_Page_077thm.jpg
e1b24f1cf640cd77c95121388c01a11b
ed5a2c00d691459dc6d086fe607c9d87abad4ef1
8118 F20101114_AACWZH cruzcasas_c_Page_035thm.jpg
e7ab5def8ee3efc5f7dda913dbde6ff8
0a6007db8af5e009234cac1e13fd33e18f67c591
34077 F20101114_AACWYT cruzcasas_c_Page_086.QC.jpg
e7c40ff706f2778b7c87ce0769c0d501
eaaa088cdbd6489e7a592192628e36bdc32118d6
70921 F20101114_AACWCA cruzcasas_c_Page_009.jpg
850617a2b194e4ceef750447fb69dec5
b62bdc53549644f159a3c8fdde453d2e381a17c5
1181 F20101114_AACWBM cruzcasas_c_Page_061.txt
458b2c3e6688377e06069bfa4604b698
63ac3de6a3e78c489ff18829dab2e516e225567e
1051968 F20101114_AACWAX cruzcasas_c_Page_005.jp2
950f546b01b4b230c4d7f2c3e519e21c
9a777c2d06700d2e8489402c4b5f18ea9924e781
2415 F20101114_AACWZI cruzcasas_c_Page_004thm.jpg
664ffd97ea3e9274efe136046e4f35f9
8fd0998c1814d9cd204c02b7edff60baafeb75c5
6225 F20101114_AACWYU cruzcasas_c_Page_075thm.jpg
02808e9eb117b35990a2dbbf58252941
1014234b647d6d3dae418a43409b56712742f511
45135 F20101114_AACWCB cruzcasas_c_Page_010.pro
33aa90f8aa9df685025e4f77866d78fb
ce1f6ed66643aebd2467e317550af4d69123ccab
47642 F20101114_AACWBN cruzcasas_c_Page_044.pro
f8bd401e0d850e751303582e45317232
d4d0d9e39a9a729d770b3a36e491767acfe6a6e9
F20101114_AACWAY cruzcasas_c_Page_080.tif
52ff6f6ad73c72859d0b4996c956cba9
2581d381d0f082a013949ad2f07dc0550bae2d02
27198 F20101114_AACWZJ cruzcasas_c_Page_051.QC.jpg
e5effe3a081f9dc94b842ec938e378eb
d53c923855412e8a7de5757d1b63a9a9d58363b7
9607 F20101114_AACWYV cruzcasas_c_Page_101.QC.jpg
e42ad0f252231f888760e9c0890af6a1
81dc6b04c900da594eeb9364a7c0339e8a84b014
24758 F20101114_AACWCC cruzcasas_c_Page_061.pro
4cb9a79a6e8cb52793bef350f6c98b3e
5d7c745385837c923a781bdcff7a7eaa455015b3
8000 F20101114_AACWBO cruzcasas_c_Page_094.pro
1e4daeda2d5508cd7c7ead68e422b035
9166d7d2e2086a7111e7dbbf85433e05ef5afdda
95358 F20101114_AACWAZ cruzcasas_c_Page_026.jp2
a6f5f3c36760ee0e7a0d47eb3e9c6005
9482eccb5c4454a78c46bd87561543482cf9a2f0
27504 F20101114_AACWZK cruzcasas_c_Page_058.QC.jpg
3e38786c9692f46da04f1f64633222af
2cfd5a887ee29bf7054696849d789e4037275476
6082 F20101114_AACWYW cruzcasas_c_Page_061thm.jpg
4f9c8247a5c51ff9f2efdf448b369fca
a30969d1b2136f8902f262c349fc7ae7ca0fab8a
8312 F20101114_AACWCD cruzcasas_c_Page_056thm.jpg
919782be7435655b5fa577310ede9843
1c5c8d43c219be2ef081d3922c6b660223fcb3cd
8868 F20101114_AACWBP cruzcasas_c_Page_021thm.jpg
c8a3288ab7459ab84673f540b661d5ab
b5718538a9944eb199b3a46adf2675b688096887
9681 F20101114_AACWZL cruzcasas_c_Page_040thm.jpg
147fcfa166271d4ba2fac4a5c046c8cb
92bb3b4e4505df3fbc7b387958bbaab83d3cece5
542 F20101114_AACWYX cruzcasas_c_Page_002thm.jpg
80b8937d475c3b8c33423f5502f2ac9b
76024074fb076968e3aafb81c6ea097a2bcc7c83
106619 F20101114_AACWCE cruzcasas_c_Page_056.jp2
8acf678e43e0e1001c44a9b1ae49d11d
4fbd27053d0940dd3ed12fc2c3656ad6de92af8d
2032 F20101114_AACWBQ cruzcasas_c_Page_049.txt
ffddb589c2636108d572552504e6c843
5e3100e488277b3a4afa1d071e469d04ee34c676
35194 F20101114_AACWYY cruzcasas_c_Page_021.QC.jpg
4e5bc5879475d34a08418786c033b331
a27a4448a70aca09aba8c3da25a8ffa8ffd3da46
6644 F20101114_AACWCF cruzcasas_c_Page_005thm.jpg
4e64fcfb7914c11369cfe73a528c6037
fbccc84218aa2656da73699b60f9840fa9f1916c
35544 F20101114_AACWBR cruzcasas_c_Page_053.QC.jpg
51d227f6f6b0fab4b214ebc36f762798
2818184fb656dfd50ff5481b1a4e2c331816aee4
4926 F20101114_AACWZM cruzcasas_c_Page_009thm.jpg
7985f979911b10196dc7215083b618b5
317e2785ed8d36757444ffd62d07ea164b28b89d
30494 F20101114_AACWYZ cruzcasas_c_Page_082.QC.jpg
05a4f1821a60e07ac5c55f8f310f5f4f
7de8a6374d15da9f95761539f4dc57fe25845a8d
F20101114_AACWCG cruzcasas_c_Page_006.tif
a6181150f762636f8f168bca94c2db1a
fe6db9628d6767929afe2559223645b52edf7a71
1999 F20101114_AACWBS cruzcasas_c_Page_035.txt
562f7de7e172c2ac620ab8f0c1756d41
2c339430425c04c3b404f4bbae302bea0b6b0589
6214 F20101114_AACWZN cruzcasas_c_Page_093thm.jpg
2efe2dda247f3fdd9b7266fed0e05054
179e2c396ef1e353f93ae4643bdbc821b2bdeb83
F20101114_AACWCH cruzcasas_c_Page_071.tif
aee5734b00154935ea922890b5e1bbfd
a7615291b5503d2ce03c0f5c131526350d4984bd
27832 F20101114_AACWBT cruzcasas_c_Page_041.QC.jpg
23ab9581c499ac584c23d67631ade890
24c2b19db903e61456c1b1467feb689b5e160e81
8273 F20101114_AACWZO cruzcasas_c_Page_050thm.jpg
7432afce150ec36e06a62d7530c5862f
ecc9124af9603694d9d6353f1cc564be40d144a3
2104 F20101114_AACWCI cruzcasas_c_Page_027.txt
c41dfbec321cf36bdc870c467993d7db
a38bd99861df09418f28e3e3985ae91fc824f313
16135 F20101114_AACWZP cruzcasas_c_Page_098.QC.jpg
3015da3ab6285b5c0db31ffb36f2cdc0
fe61d31ac9c0bd93290335f4c8fbea394d7d9017
115817 F20101114_AACWCJ cruzcasas_c_Page_070.jp2
e4371848fa37da172ecbbd388bcb24c7
382478bdef86aa1129da15828c53acedfc81fcc8
87962 F20101114_AACWBU cruzcasas_c_Page_081.jpg
4b2a30ed9fceab8f0e6041100697cf94
795cec21ae74b27778621ceca8598acc1536ea50
31023 F20101114_AACWZQ cruzcasas_c_Page_014.QC.jpg
31ad6adc209d13574a378b54472a1d0a
f4629d5d74921c011fd86084897271c84dd9430c
22083 F20101114_AACWCK cruzcasas_c_Page_107.jp2
a82656bb847b5ba89d0f888a4144828b
ceebed3cc5363e773eba860f177a102c66ad0e09
1856 F20101114_AACWBV cruzcasas_c_Page_107thm.jpg
194a9e3454f5209f0288ad881de7d2a0
8534317ed2efc9bcc1e153ec2cecb70a625e3c41
4364 F20101114_AACWZR cruzcasas_c_Page_104thm.jpg
d67cb1a3ab4a79b95c9373a04d518ff0
1930308afb40be9f04e62d94f96dec747da873c0
490 F20101114_AACWCL cruzcasas_c_Page_094.txt
4c84a021b0cd0e10066e0fe8a1c4a614
470a774f237ff8c74a91005007a7640c0701b58a
26913 F20101114_AACWBW cruzcasas_c_Page_095.QC.jpg
c291fff7b02413bc3307edb729c9594b
6da8cd6ba71f30247a54354a90ead605920c14a2
29121 F20101114_AACWZS cruzcasas_c_Page_079.QC.jpg
4a8216aee319365a2d149c77e188f10d
ccf6c51009fa7701ba4160e9945c3131064f06fe
50267 F20101114_AACWCM cruzcasas_c_Page_104.jp2
932cb7edd53717f84abb98f942cb0cfa
23d87fd87e9300a8afcda415542ae9d88697f191
113651 F20101114_AACWBX cruzcasas_c_Page_007.jpg
8842276459580b4102057fc612ae9215
e892a4f4d6533829b62027ec5a3c8fdc9774c313
31507 F20101114_AACWZT cruzcasas_c_Page_029.QC.jpg
b29e085a7a96d8348043d8939625bb1b
0a141eb0fd8a8eb98b70528c54b60d904f95e442
7282 F20101114_AACWDA cruzcasas_c_Page_081thm.jpg
4fbbb69ccfc34524c29bd6a424f231e4
16c38aae04a0f2918488c2c5276cb0deab48cadf
30735 F20101114_AACWCN cruzcasas_c_Page_030.QC.jpg
45b1ab1a5cbdc2d9c3572b7caf86126b
d38f744374a21c417205a5cb9471b2b395c669b7
45013 F20101114_AACWBY cruzcasas_c_Page_077.pro
17b248d15b5ef263b85b4fca573fd4bf
58499b83ed5b3379fb9fd14bf44d974894a3c0fe
36488 F20101114_AACWZU cruzcasas_c_Page_084.QC.jpg
800132621f3eb14580ad9f3979197fd5
edbe55a8bc60738ad8e0a8915d7799696f531193
37136 F20101114_AACWDB cruzcasas_c_Page_088.QC.jpg
6f6beb9b5b6fb8a8ab7e4740bb5f5ae1
9f25031d60bcf9ab82b76f63694256a268d40393
8224 F20101114_AACWBZ cruzcasas_c_Page_012thm.jpg
8a10939415b5f7583044cc05845c0332
8d37019e26466a6c63e2ec65d76d7f7d41154d4a
7403 F20101114_AACWZV cruzcasas_c_Page_037thm.jpg
3fc89c82f4e89609d9f5ab1f27920caa
fcdd80851386ceb64376a203af7aed6922cdc03f
1952 F20101114_AACWDC cruzcasas_c_Page_037.txt
e6b6a13fb671c74ac48d1af87fd005cd
bcb339859d154c4b823bbbc17bdd23eac759556c
91050 F20101114_AACWCO cruzcasas_c_Page_058.jp2
c86d318f52b4f7c9fe2ea2634d0946dd
65882657a93d2874440facb401d592e13a1bf9a1
29764 F20101114_AACWZW cruzcasas_c_Page_067.QC.jpg
13bdcbf3e4a5a1791a5827f28b9006cd
50dc1979171c82646d361a4a91f2195175c5416c
94479 F20101114_AACWDD cruzcasas_c_Page_027.jpg
e842b1f2a77279c6483204eafedacf2c
16682e663952cde3670b250895685d908524d9bb
17258 F20101114_AACWCP cruzcasas_c_Page_093.QC.jpg
892bc6db34c20491e18c1dbde3f5ad26
ee8db046daed3015f01f11840bedb6aa0cdcf437
36536 F20101114_AACWZX cruzcasas_c_Page_070.QC.jpg
36ac6301c8626515d9b934f4b5c14180
1203dbaea3f56f28b6c8f87e4c38306eec72d047
35117 F20101114_AACWDE cruzcasas_c_Page_055.jpg
bc1cfbb8e5386540ec1f2a2c0277dccd
80847607b344ec138322a34f05c471598eeb000c
F20101114_AACWCQ cruzcasas_c_Page_040.tif
b974f58dbe114c893471ce534e3633f5
b31686f144ced6e773cba54b9cb5ae5234a181af
37951 F20101114_AACWZY cruzcasas_c_Page_068.QC.jpg
9ee81e06398e08d4e3613e9e2e19b6c3
196c15c6281e2b28cec4a960a7fdc3d66679598c
5391 F20101114_AACWDF cruzcasas_c_Page_033.QC.jpg
468755e49914491dac40a46470af6d88
b9c3cb58746abe1a8d9133df9c4a8c924f674d74
2171 F20101114_AACWCR cruzcasas_c_Page_089thm.jpg
517becc91b32c43a4270a7e23748dfd7
0d144bfdd53408ae35de75a71f4e0185d2eb1a4b
30920 F20101114_AACWZZ cruzcasas_c_Page_102.QC.jpg
dfd9e9371f8defcbb6b4b7bfef46578c
2e9dd98a37e59bb8b3ee002dee990393e023eef2
30340 F20101114_AACWDG cruzcasas_c_Page_060.QC.jpg
5df7c14cd74d322f236c9434d53290df
bf4e76c384265ed5f48c90a074d60d4675f2fa84
F20101114_AACWCS cruzcasas_c_Page_007.tif
19df2d41bb73f2716fc1ceeb8cb4eef5
cb2e9fd7113b915e7985e0f2a18b4ad569364fe7
7890 F20101114_AACWDH cruzcasas_c_Page_029thm.jpg
38dccdd758d92d5cfb7d81a180e639b0
7138ed78e23029b7878826068f48b20635a7fe58
86843 F20101114_AACWCT cruzcasas_c_Page_017.jpg
4a45d1c4c89f0880f22e8839b532160e
9c81f86e10e8898ee35746ddfcee2f12d1c05f20
F20101114_AACWDI cruzcasas_c_Page_022.txt
ebcce1ebf973f2e6d5fde55159d30c33
b769badf7f01423431799584ad27a5139d3dc1b4
224 F20101114_AACWCU cruzcasas_c_Page_033.txt
53c4fe6714a68b43fa2ea096f01610da
4ff1664c9d5ce88b1aaed44a8bd7442db153ad3f
F20101114_AACWDJ cruzcasas_c_Page_070.tif
e75c0411bc9f17450edfad1adacec492
b7089ca5096c4e53a11f37b9493c5cba0be41c9b
102305 F20101114_AACWDK cruzcasas_c_Page_044.jp2
af125d3967de0a727f147ebc4628b66d
4c5437bd57e2e8d3fd7b3dfd99056b2de5ba3634
56011 F20101114_AACWCV cruzcasas_c_Page_069.pro
8d45c463a29ff26c49e9a42d63803c6a
d8c89b4c57bba9d087a64f5379e45dd71e5aaadc
19967 F20101114_AACWDL cruzcasas_c_Page_009.QC.jpg
4adde5835460e777abadac84baf5643a
bf283bc1d0e634ff347a59f7b0279ce2519618a6
F20101114_AACWCW cruzcasas_c_Page_031.txt
b0c6497c57bb495c1148624ea62e3404
b40f9de9c1dd83d039b9b9eac46ef3ccf85ccf6f
125394 F20101114_AACWEA cruzcasas_c_Page_106.jpg
ba31f45146b08a2c87b303f59bb55f49
f7f1f0d0cf40a1c2692c3d2b268000ce3d5cf176
71833 F20101114_AACWDM cruzcasas_c_Page_059.jp2
149bedaeb5cb3c64af05f14c1693a580
a629232bc89816a0221d06cbe70322dcb89b1b67
6132 F20101114_AACWCX cruzcasas_c_Page_091.pro
eae0ec3460d6bde5c40cd62bbca88f91
d18c3a2526e84b3e7ea2410433c33e3df4316535
8938 F20101114_AACWEB cruzcasas_c_Page_070thm.jpg
f2ad033e830c79759e21abcea3237c2b
087e01ae8396370240ed3fe8541c06b5321ec486
833673 F20101114_AACWDN cruzcasas_c_Page_078.jp2
cef81bbf56d50964a0d19af04e08d97f
29a4318b3975b3ff663475ad9109b5236c2ef8cb
F20101114_AACWCY cruzcasas_c_Page_039.tif
7b3f9f9854715749aa7f7cc20ef85fde
f17b028491d0e4be6d0bbbc3006c5d7aaa737371
6902 F20101114_AACWEC cruzcasas_c_Page_045thm.jpg
464f31b32edce0bfb5c7e90387f458e9
14960338cbf6cb3347c7d407fd81123c5441c8fe
34933 F20101114_AACWDO cruzcasas_c_Page_087.QC.jpg
0e11e351bd44b212d2356d702c6922f3
8b4ef045764aef15c66bdd7038e2050a5722d1f1
55022 F20101114_AACWCZ cruzcasas_c_Page_020.pro
76a90fa39c6be20f677975978bd12769
86fbba3a41c51c6ee2f64561ff55ea3f71721df2
35990 F20101114_AACWED cruzcasas_c_Page_100.pro
c6d463c7343a8770d5d765a1ffe276c3
97382fbb2694443cf0f7817d21b85928e3c8b32a
57503 F20101114_AACWDP cruzcasas_c_Page_007.pro
5dfe9e1e38eaf0bde642495b498d89e2
f0027d6dea2f7f2c12064abb33f9e74ecca1f425
779309 F20101114_AACWEE cruzcasas_c_Page_075.jp2
bb74ef49c947a14a7d87a41d0baef311
b83f0b95f6bdc7a418f2233fe0eaefe69b31de14
1361 F20101114_AACWDQ cruzcasas_c_Page_003.QC.jpg
cc04ac00d650ba3aabf1fe85c951e403
0874fe24e37661ca284d28a23d9b78dc11db6c60
6921 F20101114_AACWEF cruzcasas_c_Page_040.pro
9ab1f4f7e31d2a89c22f0680d5c8d6a7
41390e063a7aa06349cd9bbb7291f7f6d11dfb76
2366 F20101114_AACWDR cruzcasas_c_Page_103.txt
15f7218f76054ae0e4ad5f473908f76a
62c735173e48de3c2741c4d9d9fe5a84b3e9a2f7
F20101114_AACWEG cruzcasas_c_Page_058.tif
36c8ae933d230dccfc88d81ac45210fa
709e802dbb6fa6c0b0511d075f1d69e7f23ff50e
171 F20101114_AACWDS cruzcasas_c_Page_093.txt
7a7fe4246947bd64f2467efc2f9ec2b2
5a7380a235c7cef1a0946ed271006540e3114ebf
1026 F20101114_AACWEH cruzcasas_c_Page_038.txt
ecdbabf6ae9ea83e41898fa6742654e3
b57761717729407a50a6593827f0196099990745
111001 F20101114_AACWDT cruzcasas_c_Page_024.jp2
4bb356fc99deece8e9452776c82ace28
858e72d1527f539ea2b736e345051cba139928b7
32899 F20101114_AACWEI cruzcasas_c_Page_031.QC.jpg
741c9b25a10e7606ed00297fbc2eff90
b3ea10ff1940ebdcb3f1e309348864fc2a2bfc8d
1871 F20101114_AACWDU cruzcasas_c_Page_060.txt
6e65365166178aa2c20554399301c32a
82aa9cc8337b9a985457e339a768da625dd3ea8b
F20101114_AACWEJ cruzcasas_c_Page_016.tif
148e3d46fbf009aa682d5ccc47ea0ec8
de33cc3c66b1154614e467bd3a81b035ed1cfd8d
94304 F20101114_AACWDV cruzcasas_c_Page_103.jpg
ed0309c21ac45d65894cd62ad2e80883
88eec9e6a84bb2420af9afe3e48a7ee66291dcbd
22050 F20101114_AACWEK cruzcasas_c_Page_038.pro
a6c59bf18b1703bef3787efede0d8d75
2d2cb645cc5d7f563847bcbfc25a5e005d7a15f0
856684 F20101114_AACWEL cruzcasas_c_Page_018.jp2
84213be917e5e36ccb527fab82413cfa
9d4c590694598ebffab7f7b2843eb79198cd69be
103764 F20101114_AACWDW cruzcasas_c_Page_035.jp2
11de772e342e674ea73498f33fc3ba88
8034a8848a50b6d020f635341dbff3840bd4c9f9
4572 F20101114_AACWEM cruzcasas_c_Page_090thm.jpg
d1d1cb361914d0dccdde03cf73447c56
ca3111277b8c7e8ef5267426dd9f754f2e1abbff
2094 F20101114_AACWDX cruzcasas_c_Page_086.txt
72b4de891b25295c8099a84b7fc811c9
c348ba8586ef33e2864db1ddae8404fa3dfdd28e
63733 F20101114_AACWFA cruzcasas_c_Page_105.pro
78d16ed730d00344767a4bde4241e81f
8f83cb4e26cb0b724882330d019d0f131eb1a5a8
7825 F20101114_AACWEN cruzcasas_c_Page_046thm.jpg
a4c5c894a033af6d260eaeb72db84a81
370dddd33bea305e13347d506b13c2e999c73c82
8515 F20101114_AACWDY cruzcasas_c_Page_089.QC.jpg
adefce285f488e4913fc6cbcdf90d4ff
5dd51f1f89a2fe321d5811f05ab6f44182fddff8
F20101114_AACWFB cruzcasas_c_Page_104.tif
698849998aaa61556da944c650e4ceb0
ead8566f63770a3530d46e883e695609b7e6dcf8
21556 F20101114_AACWEO cruzcasas_c_Page_011.QC.jpg
6aa4ca3303724b013c2212d58dcbfc16
f09b199162d802f15ffba4ec46800b176dd3bc74
76241 F20101114_AACWDZ cruzcasas_c_Page_025.jp2
36ce8e6954c2178ece89e550905071cb
1fabff8457978fcdba166225296b13e3b36f9a19
33149 F20101114_AACWFC cruzcasas_c_Page_076.QC.jpg
94603b2db5f1a73483d1b376c5971aa3
03818a594b10c1a1437d6fbbf9533e0ea97676e4
47947 F20101114_AACWEP cruzcasas_c_Page_057.pro
7098a9df600789cf9b0b64a9b65a599e
b217bb821ea180c67b4eb79c9a035b06a6473d6a
13271 F20101114_AACWFD cruzcasas_c_Page_083.QC.jpg
5e13de9cffe10d377b008d3bc72543f7
11fbc153c8af75eb623e5848cb2db4d4ba0d961e
F20101114_AACWEQ cruzcasas_c_Page_057.tif
485b16b5e6b4a876052ba58b75b9161c
42f7c09881a6621fcf6a1e50b5bbc309b3102bbc
25199 F20101114_AACWFE cruzcasas_c_Page_001.jp2
734a6e6b54e55e8592f1d2a9089916a6
16448c7844654dfa06275b35b0319080c782be9b
F20101114_AACWER cruzcasas_c_Page_061.tif
e63a2a954b197c7092fc65ca226404ad
57382efd76d76296ebed97e0f41f6d6e557e7706
32995 F20101114_AACWFF cruzcasas_c_Page_025.pro
53e8edaebd3880e933c95bd622ad41ba
3ffe7137aa415b28ab059f97de9623dc6a0ada6d
F20101114_AACWES cruzcasas_c_Page_051.tif
655a468bfd6ad7ed6279897debe700b0
894bab5b51c2db25c3809aa46ae9f4e62ffacd7c
25110 F20101114_AACWFG cruzcasas_c_Page_074.QC.jpg
119c61d27fcaee8bd538867e5c6becc8
38b9af92bf7a3e3c1ea141a0f8a06dc9e7e8e6a0
106914 F20101114_AACWET cruzcasas_c_Page_032.jpg
035a32eaa3b5ea54dbdc1865630842ff
c21ebb4747cad02eb22c473e1bf4a5a5a599b96b
46353 F20101114_AACWFH cruzcasas_c_Page_052.pro
1fe16a474ef23e5e8438ef87e3a18f05
5a1e167e97daf3fbad2610a4e0d14ee3a94ff0bf
46220 F20101114_AACWEU cruzcasas_c_Page_062.pro
b1e8df4c495e41566d84298010848176
5e5f049da4196560b1894be9ecd02c6ea1561f6f
109200 F20101114_AACWFI cruzcasas_c_Page_049.jp2
6f9f40db3cb514385cb9bd7c2f876412
32a4d2ddc986fdd5d7d8ba6197c4e54bcc417fd3
9079 F20101114_AACWEV cruzcasas_c_Page_085thm.jpg
3b9d22220223e7faae03d444c80982fd
7e222041e481ace33ba10c0bb7e8f797542bd2a0
F20101114_AACWFJ cruzcasas_c_Page_076.tif
18b4b19cd9391f691edae62eb390b714
c9d0b0913e9c5c3ba2f7d33d08352e5a94a1b2f6
100542 F20101114_AACWEW cruzcasas_c_Page_071.jp2
b4c86c7c94153b920a64503598e06a55
15aff3e4090633f0c0e8a629b05ddcbb095f646a
89667 F20101114_AACWFK cruzcasas_c_Page_026.jpg
5799daea482a56423749a9856b23f43a
d59df0160ed87685a7caa1ab56f6321dbeff1b2a
14963 F20101114_AACWFL cruzcasas_c_Page_033.jp2
29a737fd11dae6d28edbf6fb4b935e17
eb9cb47f4b85e8847ffff68485ca60ad319f3c63
8223 F20101114_AACWEX cruzcasas_c_Page_076thm.jpg
d5a0ead65c372037b83430258650c27c
88a1e365452afc46e61410be6cc07ac67ea8455c
33143 F20101114_AACWGA cruzcasas_c_Page_007.QC.jpg
8e0decdb244594f8299aa8ed550dadde
1cf374c03aae63ae9ff29c0efa12f2abadc9a9e5
F20101114_AACWFM cruzcasas_c_Page_064.tif
35424c3b7670f0275f09fd0c662d7e8b
fb9d2756ffcb2b05c90a9c18ba2d310c37f6b1df
33466 F20101114_AACWEY cruzcasas_c_Page_009.pro
34ed7d0d7bc1c8f5409238746c5c082a
709fc26bf35491e0db30f139a1f96d5f1944f957
2312 F20101114_AACWGB cruzcasas_c_Page_064.txt
c2353de40c3494f6343845e1137270ae
00e561883974c1c797975a8aae2c753ed54da5a1
108642 F20101114_AACWFN cruzcasas_c_Page_087.jpg
5dc2de5cb41e706a905fb9ff62d20701
1dbd8e825af26ec18ec0d61de8fedf9f86ef3740
118375 F20101114_AACWEZ cruzcasas_c_Page_036.jp2
ce7afdf4c8367d216da7a4410a6171ce
73d73e397b50599c9f77f639fbcfd5121b5cc68d
61267 F20101114_AACVZT cruzcasas_c_Page_008.pro
46ae1d2e9681c5c1217662b8f4444fda
d27d28436773b2e9a45475fa6ffb6407da447066
12503 F20101114_AACWGC cruzcasas_c_Page_004.pro
7c06bcf9278bc441025ab0eab7d7d267
c3997f260edabb6dbe985d50eda08c6138ce84f1
F20101114_AACWFO cruzcasas_c_Page_005.tif
387953733a32db813cb086f2bc350a70
f1316d5f7dc3bfc41bb1154d5b756a9f05f6886f
94498 F20101114_AACVZU cruzcasas_c_Page_062.jpg
e16b7cf76193a52daad6b1434e1f7fe0
39954da1f4d450e2306d1f6fbf87c506888d5183
F20101114_AACWGD cruzcasas_c_Page_001.tif
b4af4e8d26d45116cdc46306ea1a30fe
88f15f1ac133abb1df00bc11a2b17143aeef4400
25500 F20101114_AACWFP cruzcasas_c_Page_080.QC.jpg
f5fdfee316bfe3aa16d649e83052c02a
1977ba02c5184b52d2d5031058c0fc67e6487eec
9057 F20101114_AACVZV cruzcasas_c_Page_054thm.jpg
c87284b5301e662e9b5c6d738130cd28
c6ce995e029cabb645461bba4d75ef5cdce20b73
7012 F20101114_AACWGE cruzcasas_c_Page_080thm.jpg
39910c84815355de15a59e6e2cfe377e
778540226b44075bf4c053a3a9d2a7927c0d3118
8285 F20101114_AACWFQ cruzcasas_c_Page_031thm.jpg
3891122444e40f2e52c3dfb8cd737db9
cc5c83e744376f3b4a6e6f4ee3c88ce095238b76
44883 F20101114_AACVZW cruzcasas_c_Page_090.jpg
f07bf873d8876c46f7148e500a7b403d
22c8147502ab95bcd4e119ef7513f07fbe86b847
700746 F20101114_AACWGF cruzcasas_c_Page_092.jp2
7ea33f4359c89ca6d6498f25a08165ef
841dd901877ecc09fc08343a851abcd32d7f6ea9
2093 F20101114_AACWFR cruzcasas_c_Page_087.txt
b9c7999fe2e56449d3968a42a5829b81
fb284545e59ed8936bd70887cec2dcf176410ade
678 F20101114_AACVZX cruzcasas_c_Page_066.txt
db55fabc3d34d4bba5a4e7fabbc3d027
dcc9aed128daab724c8ed019e5e1cb9510b1060d
72961 F20101114_AACWGG cruzcasas_c_Page_028.jpg
192137bc990cbc5775c0eaf880070f21
0e7186e5c3dba7e77b1f5204f65ca873d1dc4b93
530793 F20101114_AACWFS cruzcasas_c_Page_093.jp2
4ce7310b97c91c449e017e58ab9715bc
4c581a089a8431d7a6c58b68f0eb56a67754b3b0
F20101114_AACVZY cruzcasas_c_Page_028.tif
cd0ce651f3941be21521ac068ac12c4d
6d597c30137a1740f9fdff6680b32f6934b0afc8
86550 F20101114_AACWGH cruzcasas_c_Page_018.jpg
7942be960f8cf017b738e24dae50549d
dde17f30980a85126f7229db00329c82dacb3736
111182 F20101114_AACWFT cruzcasas_c_Page_039.jpg
c9299bd82dc9bd290293b02cf1425217
cef150eb4698487f2facd85fa35ece8c1d6e3082
28419 F20101114_AACVZZ cruzcasas_c_Page_078.pro
77bbfb47494dcf53fd758e84cbaed678
7f6a98b55522a2646aac8a51811d7296e43c3c3c
1815 F20101114_AACWGI cruzcasas_c_Page_067.txt
f53481252084aa844462500864a2d024
6afb460f7322711cfa66eeb967a95a2849edf3bb
94235 F20101114_AACWFU cruzcasas_c_Page_051.jp2
549b7d1402fc01141ab3defbb3a45a40
af54d547df802196f8f8b5d0fd61e7be9f88eea4
33817 F20101114_AACWGJ cruzcasas_c_Page_039.QC.jpg
1bfa385bda1051059ea2d2e149ca4648
5aa657251eebbd08ea76624acd14dec3508b9f24
27626 F20101114_AACWFV cruzcasas_c_Page_101.jpg
cd0926fb8e2a6871731b62015a984f45
d91ce4f8637d9f60db5d292bbefb0a6681d58c60
1608 F20101114_AACWGK cruzcasas_c_Page_080.txt
cf5f335101f3ca49d4afdfc26c65850e
440423b45ce678fc1403859e252b0d2e77d87aed
48136 F20101114_AACWFW cruzcasas_c_Page_030.pro
413827c70360bd15ba5796709bfb39df
348abc1657a729cd9e8770f534c4e454f5234109
55744 F20101114_AACWGL cruzcasas_c_Page_084.pro
e0f306fa386b065e2b48cd074abc5fcf
89fe4f6987d9a2d0336c6495f4a8e7febaacb6c5
56108 F20101114_AACWFX cruzcasas_c_Page_019.pro
749d77a77c436cf676f3d412e0a0fb2f
fc2b9939463179ea8fbcc8ca059e9eeffac8ced0
72464 F20101114_AACWHA cruzcasas_c_Page_095.jpg
6763ab0b661933fce5e9970bc31a86fd
baaf98e56a7b364fe1f221113e0e53d0336ed28b
9017 F20101114_AACWGM cruzcasas_c_Page_064thm.jpg
563c1bd52fd277c424ca22c9acb93671
122e4c41b5f90c031704a2aa18d2a821eac04619
F20101114_AACWHB cruzcasas_c_Page_074.tif
b3a1e6b7f912be1f879dd693d3c26c27
b2b176730d4dbff3172647b625c1cab2e3a94bc8
2851 F20101114_AACWGN cruzcasas_c_Page_005.txt
47bbe5b8005ed42ccd73126356e19351
6d0e71f22e77f1bce80d096cca6551176d377fee
28124 F20101114_AACWFY cruzcasas_c_Page_037.QC.jpg
7ad9d79b9a5abda3ca78601e7e976930
cd0c81a4d637c06a67acb350f5360b4fa57fe682
100005 F20101114_AACWHC cruzcasas_c_Page_076.jpg
15d0c27fe679e98db03873009627b872
45de183cb4736431ecabc60e49876040fa6e8091
40583 F20101114_AACWGO cruzcasas_c_Page_098.pro
3a643f75e38e73a7d4815ba7fa01b9be
5681382be4de49776c5543bb3b52a105d0cfeaf1
38165 F20101114_AACWFZ cruzcasas_c_Page_072.pro
3ba86bfd0c8837d57b728df1c8cabb84
5657d866ebecd20c8e30e7970bba15faed31232a
49976 F20101114_AACWHD cruzcasas_c_Page_014.pro
454061b5ffe36adab44f11c2d76bfb25
b2ebd18e2a329b868be1c919a68d7f614a48800b
55517 F20101114_AACWGP cruzcasas_c_Page_100.jp2
993bf3288bd4fa70b49887925b87c21f
26da96e9f84b335ad89532f70dab037142f2c643
24096 F20101114_AACWHE cruzcasas_c_Page_078.QC.jpg
93de1cc28bc5eb78f5a3c5892b0aadc3
db652ff42f18e3d7e52971a771b88f35e14c90fb
47371 F20101114_AACWGQ cruzcasas_c_Page_027.pro
0dc37484e089eebd124c6b389010060f
51181233ae49e026bdb50dddd43911e269915d67
99288 F20101114_AACWHF cruzcasas_c_Page_027.jp2
c865c7db545248b9faeb424bdad0cbe6
265c3c818c824fa23719c93268f25ca226d3e6a6
F20101114_AACWGR cruzcasas_c_Page_096.tif
a1660c602058c11d933f3ce740096ea0
c8f2b6004e0c94b9aba83eafcdd4e930ebfd26b1
1893 F20101114_AACWHG cruzcasas_c_Page_044.txt
85274c7140978c9aac9a1a9f218ff968
e842cd0546505b6663225e9cf70752b5fc4de6e0
80022 F20101114_AACWGS cruzcasas_c_Page_091.jpg
99b3a1d5182dc8906576379cdeffcdc8
9a5c3182ed2b731c719d4e3a5fbac34b1ad797db
116675 F20101114_AACWHH cruzcasas_c_Page_031.jp2
ac89a084a07fe71763090076aadac3eb
ecd955cc05b1d7b1165ff754bdb0581e56fcc845
2514 F20101114_AACWGT cruzcasas_c_Page_096thm.jpg
8adb69add7b0bf536ed4cd454d65a44b
e1b47efd48ceb5ef848ea0fdc368ae15f087ae6f
896 F20101114_AACWHI cruzcasas_c_Page_047.txt
732e83079a337d04bd698d3490e34dcf
fc3865ca272b3aa3a777dc0c4b4980d46583da9c
105611 F20101114_AACWGU cruzcasas_c_Page_024.jpg
47f7ef3d111a195cfd917127bf61c9fe
6f25684733b374912f42c1343016f4a6d93450e3
35916 F20101114_AACWHJ cruzcasas_c_Page_018.pro
1454ce6c29ab9bd541d605bdff98af3a
cf395c87c87cc0afd2b66202f8412c7df23e0cbe
112412 F20101114_AACWGV cruzcasas_c_Page_036.jpg
f5a81bffd31d9c17625daa4204b297a4
541648cb1d16a5a4212162619a1ca848d0cb1579
55758 F20101114_AACWHK cruzcasas_c_Page_085.pro
31e27f304340e01f84e4d3dde775fa80
e723d02ff37872eb0abf901224dc16864335faa8
126 F20101114_AACWGW cruzcasas_c_Page_096.txt
6cc90e01e12810c686a5dc2fe55f9939
4db173dd69441922069c555a805270e26ee9fa41
9448 F20101114_AACWGX cruzcasas_c_Page_039thm.jpg
fec819c991ad062979a29f8b3619fbbe
3b5ef30ca5f60a92a5d6876e703788e83a69ab92
8183 F20101114_AACWHL cruzcasas_c_Page_052thm.jpg
7420373e57060e71d160cf400fe281c5
d7c528181b44cc7140e2d57807237024f2ccc970
F20101114_AACWGY cruzcasas_c_Page_024.tif
a504edb2c08730f575f69b3547014c36
897f199ed0b9c54810a5eceb539542d50e49b2bd
32734 F20101114_AACWIA cruzcasas_c_Page_032.QC.jpg
3d55802d148b8c5f5a54944d772cdcde
b702ee5f2db682bca9edf8d2fd9014d44412ea9a
6856 F20101114_AACWHM cruzcasas_c_Page_051thm.jpg
2ba5341d789d9dbf3ee2abebd38db84b
0a110b9c616ef113a9e4ff97e9b2d2c772976e32
2570 F20101114_AACWIB cruzcasas_c_Page_024.txt
15008ccce1cc3751bb3d5eda2904c621
f8f4de4bcf157580fae0088a76cf8293f78de81e
86928 F20101114_AACWHN cruzcasas_c_Page_058.jpg
63a8489fdd141714eaae77870a42425a
824caa0de2db5109ecc21b2904bde97ec0d3f5b3
F20101114_AACWGZ cruzcasas_c_Page_053.tif
14554c038ff9412faedaef8e38a30c81
f666fc28fd43dcfdf9442ee993940bafd4eaf074
1964 F20101114_AACWIC cruzcasas_c_Page_001thm.jpg
ea3a20d08a07fe9a6cf1cb75e674f02d
0f4ee46e68ec0a59215f129381617ad5aebd3bdc
4872 F20101114_AACWHO cruzcasas_c_Page_003.jpg
6c745e20f5e4338f541c1fcbba471a6c
98bc698b94c336fa7b5ddbe8f38f1927714f298d
84228 F20101114_AACWID cruzcasas_c_Page_102.jp2
7345eae1e0e0e8fdd9392c3167fc8628
8cdde67a2276dd98e6ebba87bbf4795bd0ec46ce
50710 F20101114_AACWHP cruzcasas_c_Page_068.pro
22f6fd82bb4e89b6813e812f59f80a84
c1b26eea9b9d798387b54702249f00855fc3b756
5820 F20101114_AACWIE cruzcasas_c_Page_015.QC.jpg
30ee5d51c9aa4f91b9374fd769aa975e
44dabcc34a594053187b94a511d24f9b8335e262
7497 F20101114_AACWHQ cruzcasas_c_Page_042thm.jpg
c74e93821b9c555c87ab5f0eb380977a
947171386d316c580ca4d9663ebe7bf82ae9f6f6
29679 F20101114_AACWIF cruzcasas_c_Page_052.QC.jpg
f8418dfca7c0b428a5cd3550546e6b54
9a1f73e35c963e1c090a30adf6c9abc517135d0a
618 F20101114_AACWHR cruzcasas_c_Page_003thm.jpg
315e84e8d8afff2f95bcea1d44c755d9
16865a1cc8348b8fae1461766bf04d5703f3e104
53228 F20101114_AACWIG cruzcasas_c_Page_087.pro
bf12071c268c9284c280034aa0231010
a4c3a94529753e791c2c0d2a9b51f69ad8265ae3
52090 F20101114_AACWHS cruzcasas_c_Page_103.pro
18fd7a819015b5e1a592d62b90a1b2ec
23082c562b6104fde831f080f4e545fe5f39a7ec
8121 F20101114_AACWIH cruzcasas_c_Page_107.pro
17f5c5d9984e00aada69d0f030b36521
be3dc52a9953de9fad9a16fad2a3c8aa9f012df9
F20101114_AACWHT cruzcasas_c_Page_057.txt
dea53199271a6b724fbe663de9a0c536
ad9df6bae538d54cf25110cde8d1666a78ad17b0
97566 F20101114_AACWII cruzcasas_c_Page_047.jpg
33d79c5af6e66fa4674f64ff32855941
525df45438bfb15d8c6b03ee6d244e8c81b8e78b
350 F20101114_AACWHU cruzcasas_c_Page_039.txt
b5f324eec0b51ca109dda0fbfb523fbd
3889e3aff5a8f9612ce79ee5402fa6554f028a2e
7666 F20101114_AACWIJ cruzcasas_c_Page_057thm.jpg
86d2f95ac5245794f6687b7bc1937ed7
c59434be265755e83d4b193f521055a171f9e759
2124 F20101114_AACWHV cruzcasas_c_Page_048.txt
0aac46bb444294a59e9d8e4e1178e4cf
d759b25de7900cb6bc8d67680b68f28f1a639256
50264 F20101114_AACWIK cruzcasas_c_Page_076.pro
a2f90c368d1972b0f294433909c53e49
488fa187afbeef947c0c7db82e3f12b362840524
2899 F20101114_AACWHW cruzcasas_c_Page_101thm.jpg
6d3828b77ebbf9a87361a894aa4a8bf2
410ce97f182a6a792b5cc96bbf8e2261e37d49c0
35000 F20101114_AACWIL cruzcasas_c_Page_016.QC.jpg
e52c1551090d3139513babde2f737de4
2fdc7dca83ac97c61d34cb1ded5db6cfa224ac73
707 F20101114_AACWHX cruzcasas_c_Page_083.txt
fd9aa7c36631ecae21c88a5a8f581802
3f3e924fc199a9f59edaf9dc29a29a8c8142e63b
29979 F20101114_AACWJA cruzcasas_c_Page_004.jpg
799e52cd36cbcca8bb607e71167f9d8f
3437a1ba23ed46867518fa95d2fe276323724179
64261 F20101114_AACWIM cruzcasas_c_Page_094.jpg
8ef2e70b59ddc30795671db230cc5cf9
fc3183a9c4c8d34766ddecc382fae0a460c7c2fd
9098 F20101114_AACWHY cruzcasas_c_Page_105thm.jpg
da00b0aa437835391833fcd513191454
1862e4768619834510a541751a6106c4641570d3
114401 F20101114_AACWJB cruzcasas_c_Page_005.jpg
f765794ed69986d3d050f03c1032c0ef
51902b42ae18ff7310ccc5cc1a729174e0713d77
38210 F20101114_AACWIN cruzcasas_c_Page_083.jpg
b8f303fcad65956685beb005d2cf5d08
6ded9a3d80c58991bda91a18297e53c38a9795f7
F20101114_AACWHZ cruzcasas_c_Page_041.tif
08111859f9b6758ed35bda62139b5a78
7eb4079b92395d224f506ae4f88ed515059a8352
42551 F20101114_AACWJC cruzcasas_c_Page_006.jpg
d6fa6497034ea12a5caa7259b6e90bcf
706a61884da5e01bb9cc6547f3e70f755d981b06
93146 F20101114_AACWIO cruzcasas_c_Page_017.jp2
2386253862d73b8ec3ebfe871601dfaf
70ca8795db97a8b44e5f8d7049bfc9cbb87d00b3
123397 F20101114_AACWJD cruzcasas_c_Page_008.jpg
be36c2936623fb332708ded68e613957
94112338d35036e9faf4856caccd0087466278b6
1956 F20101114_AACWIP cruzcasas_c_Page_063.txt
464595261f3eca1763e2a0b7669cff3e
a89929abaec3398f81c16900d807f810aabbd6d9
95280 F20101114_AACWJE cruzcasas_c_Page_010.jpg
872be38d3cbb1e07e43026a913e3fe40
7f47b2a23504a7a233adab1e5d8b335ae1c3743e
30732 F20101114_AACWIQ cruzcasas_c_Page_011.pro
6d60e9f43940bfd8cc16f8f70aaf4497
73b8b160dc7653e7f416e3a1a4054b129191905d
65733 F20101114_AACWJF cruzcasas_c_Page_011.jpg
606438b912fa34d94e2489702e963168
3399afac2274877185835a4078308a5d8cf1e86b
2932 F20101114_AACWIR cruzcasas_c_Page_006thm.jpg
c7ac151d043efe9a72662f71ffda27a6
2f8614e4f1a12120eba17aa1370d29e8ec34a9d4
104475 F20101114_AACWJG cruzcasas_c_Page_012.jpg
49a29369549a3b9e5d2cc3810288cd45
abce0e70de370dc9ec488ddd7719b5bb0d065d74
1051823 F20101114_AACWIS cruzcasas_c_Page_040.jp2
b575d713a9175492822fa8cf475629f2
4ffff5dd683386f32d0710563b62bb2a30cbafac
110384 F20101114_AACWJH cruzcasas_c_Page_013.jpg
07b26ac3f6f44ce9fa880cfd2dd97245
d0b4b5628797b112e4be54cb64f06b568c63d05c
7948 F20101114_AACWIT cruzcasas_c_Page_062thm.jpg
6c958aa968ccad68726eafbe91c7bd8c
1d556c98c49aea7623eadd3e1387629811e5cc50
99871 F20101114_AACWJI cruzcasas_c_Page_014.jpg
068444eecfe2a460ae9e239771b21fc5
4efc976fa79ca6401674ba5ae2ff5002af888411
55226 F20101114_AACWIU cruzcasas_c_Page_070.pro
fbcc3015846c02b5e6848d7c18bb0bcf
b81e2a5bd15505159c4e369864e5d619a169c2c6
15948 F20101114_AACWJJ cruzcasas_c_Page_015.jpg
4a057d947d0b0027b6ae506b99f87836
5a71e35739f30b27ca8e83ad4f26a988a24ec4c5
126131 F20101114_AACWIV UFE0021137_00001.mets
17e2a672084eacc4376f300ec1073215
61c845de60acbe02e9f2c5649b073f31a7b52944
106196 F20101114_AACWJK cruzcasas_c_Page_016.jpg
a74862d8e5c5ac706cfc85db6f2fc505
75fe14b86d285f01c7102db4d5e5050213121b43
114024 F20101114_AACWJL cruzcasas_c_Page_019.jpg
9ef49c167c532b4a55ac0b26b41f3d79
6918b6726ecafdcbec371f720cd0fe8741cca870
109502 F20101114_AACWJM cruzcasas_c_Page_020.jpg
308ac1353ca0b32097d7291054c70b43
e254352804ae51ab3a16fb54f9b780612e411647
27702 F20101114_AACWIY cruzcasas_c_Page_001.jpg
bec351058710f365f93d33299b14b9f8
75d509bd914be38de9c7d29ad399ba1bf4b67d6a
87706 F20101114_AACWKA cruzcasas_c_Page_042.jpg
a00e73f759ebaa1ed9dc02eb1aaef4da
293b6487a22b37f1e1d4a61db08b7ce4a37e1af8
106946 F20101114_AACWJN cruzcasas_c_Page_022.jpg
45b6c7d271b9a4f08267d61f66e9fbbc
73dd7b06e98de287294b407126ce84086390efc5
4277 F20101114_AACWIZ cruzcasas_c_Page_002.jpg
ace49af7e2ab997ec3cf1a1eaf470ca7
1dc43b36ffddbf0b09b881f0c6050692568b69df
97864 F20101114_AACWKB cruzcasas_c_Page_044.jpg
e2c38a0fcb7921c438fabfbb682efd32
c77dc8083f6d77a7d3e700dd7c1e86530c4c133d
93776 F20101114_AACWJO cruzcasas_c_Page_023.jpg
c9704693e3e9ccbcd3459e1a2e75b290
042af6301ce8e53127b8eb7ae91593815d7600c2
64643 F20101114_AACWKC cruzcasas_c_Page_045.jpg
53102573c5ca2a0928cbdafb4b7065a2
ca8f46f4d1f0895604e6fd468dd361074b407d2e
71609 F20101114_AACWJP cruzcasas_c_Page_025.jpg
1e1b61e9a14277d21d964fa2f6dfeacf
f2bdc9624d68b95d1f71ecd84a043b6b61d1185a
87322 F20101114_AACWKD cruzcasas_c_Page_046.jpg
6ab57f64a0b7f8dc5ceba4364117b482
e061637b189876a32123b27bb238146bf1e625b9
96671 F20101114_AACWJQ cruzcasas_c_Page_029.jpg
be3865dc1159b65180b81d5e45ef1ae4
78f361466d4e9c54187158eb7f93afd5af5d8243
86027 F20101114_AACWKE cruzcasas_c_Page_048.jpg
4c0d665781ddd5f4ad9cc65311b97fe4
afa73949a6300b71103fba43c7762bf84fcaefc8
94314 F20101114_AACWJR cruzcasas_c_Page_030.jpg
4cf48b921272585055e63a6bf963c61a
e4fecf82d2f61e4ecbb0c5884b324c13fa01198c
102999 F20101114_AACWKF cruzcasas_c_Page_049.jpg
a4885aaff85fc0671dcedeb1d6562fe6
123dad16d3c83d6d07bb8a44b7937d50a6b61777
110045 F20101114_AACWJS cruzcasas_c_Page_031.jpg
04ede5a34f9292005bfb0f9065515724
a243791c77274d1a6c1afdac64d2b9d059be2d6a
98440 F20101114_AACWKG cruzcasas_c_Page_050.jpg
cf837fe4587b4f88a81cdf7d065c99fc
76f7a13ecdaf175f8ffa163589fdec6790b26491
14350 F20101114_AACWJT cruzcasas_c_Page_033.jpg
3b49920078265dd9092f2259faff55c9
9dff5f520a5bf8fe54a058e0f3570cdfb12c09fa
89372 F20101114_AACWKH cruzcasas_c_Page_051.jpg
a95db2e9634fb676f45872032bc0333d
090f79964fcd1292bcb1c2f387512b83238cb164
104466 F20101114_AACWJU cruzcasas_c_Page_034.jpg
4774813fb6258eb50cf1fe9ff16caddb
15fcc464154e96ff23813e461c2129e92cc49d1b
94002 F20101114_AACWKI cruzcasas_c_Page_052.jpg
8e3c78e1ac990cb408f6cb7de47ee93d
131c636c8bc59d35b0595ca0885c1f27cf6c6fb8
97125 F20101114_AACWJV cruzcasas_c_Page_035.jpg
a056c2d177282cc926630a4d99242075
4e7c07dae9bfeecee1909ba894553e7818e425da
110674 F20101114_AACWKJ cruzcasas_c_Page_053.jpg
0e8501a052a2fa39955c2fe754d101a3
0c77bdb3be0b070af728e30dcda25cb04d8977bb
92271 F20101114_AACWJW cruzcasas_c_Page_037.jpg
6fb8707d05b34f459320ef0f366f0df9
95949be1614cbd59187d72b10281607f07e2f55a
111171 F20101114_AACWKK cruzcasas_c_Page_054.jpg
8508d073c111d486bc2e2f5e876abbf0
97fcaa0cca931c46fde385ab01fe837026ebd235
79482 F20101114_AACWJX cruzcasas_c_Page_038.jpg
d5b9edc854d4aec0d77e78243595902b
29e1907c1f36ee50f120ff65ed8c62b5fab621e0
101639 F20101114_AACWKL cruzcasas_c_Page_056.jpg
6715556efc4aa700502f2c1b0386a760
dd1879baa399765458f9d4c76e50a0f3f0f808fd
115829 F20101114_AACWJY cruzcasas_c_Page_040.jpg
e9eeaefcd1eff1c45f4caa75079cdb06
acdcb0e81a9e1fa99240bf5b461da15fa231d803
69627 F20101114_AACWLA cruzcasas_c_Page_073.jpg
ee08d0efa8fba5e93b5ba8bb1ae2760d
74dc40f08a5f05fcb8b18ac79ac10bce81564365
86189 F20101114_AACWKM cruzcasas_c_Page_057.jpg
fc58cd0e72322b63b65f7e55ab7c97a4
ef9f933542e02b3137f9d9bdfc5ccfc0c097b072
83963 F20101114_AACWJZ cruzcasas_c_Page_041.jpg
56cac0876a58c073fa984d6e9c39f510
701ae7d36952a0915c39780de0a4e0909a7578ea
74518 F20101114_AACWLB cruzcasas_c_Page_074.jpg
0bd3d03ef0aec2a7c98302048c5509b7
56ffd96796952815609fba130ef415e994452d12
70013 F20101114_AACWKN cruzcasas_c_Page_059.jpg
690df8b0733f946fc6e9a4c14f57a04e
737d894fab0c059a0314add8448c28d7a9236f96
73406 F20101114_AACWLC cruzcasas_c_Page_075.jpg
4e4d8bcff862ac2d98a3a80d6ff30f3f
25fadbca78ad41710f8449ce44e7a4899357a8b1
91041 F20101114_AACWKO cruzcasas_c_Page_060.jpg
fc526a3fee95e898a8db0dc13991e3d5
dfaef168fc73a4fa59b2f38fa8c3ff83ea7dbb55
74496 F20101114_AACWLD cruzcasas_c_Page_078.jpg
ed5685d8f2a7fa8b1e9283e72f25b30c
f3f3dc0035ed8d607258e8b559f6dd370ba753f1
67098 F20101114_AACWKP cruzcasas_c_Page_061.jpg
9eb22bd7ec40c83aa41144b0ceef7066
80e1c4641c3e649b69d3ccd21a84c924a1f97d86
86904 F20101114_AACWLE cruzcasas_c_Page_079.jpg
4cf0acc9b7b7d51548637bf3e380a80c
6ebf1303c0b9c0bdc814ca45f4f429f8c93ba585
92520 F20101114_AACWKQ cruzcasas_c_Page_063.jpg
d7ca267cd2d82207415900d52d457fa0
6e34314f76f05e3971cbb4735f735c489ff7bfd8
76110 F20101114_AACWLF cruzcasas_c_Page_080.jpg
c25ea2fef6dbd2e498bcf2d1cb046ea7
606c48517d51ffcd18cdb1ab207e1ee6f0aefa06
110971 F20101114_AACWKR cruzcasas_c_Page_064.jpg
885493e57bf25d6ac9d69d7452a3617a
5b0ea653b713304b174ba1652b1640ff1e5daa62
92204 F20101114_AACWLG cruzcasas_c_Page_082.jpg
6422c6e334c370e79ae5581e597eb3ca
3c0a77c75a527c0ce2b7eb16b20112e35ff633eb
104963 F20101114_AACWKS cruzcasas_c_Page_065.jpg
04d6eff6b856771b9379177883bf73ea
a6b258aab3f622da75b0fc0ae51c52c063f24ac8
113000 F20101114_AACWLH cruzcasas_c_Page_084.jpg
b438944038f40368bc01994718230f8a
0ad1c255ae7f0ae111a2ece17a4751500e1e6bdc
33811 F20101114_AACWKT cruzcasas_c_Page_066.jpg
b03f92a2ff4f869e9baad7409c17b1a4
eb72a5d30a436be7d8fa38f9d380d90be4114781
109914 F20101114_AACWLI cruzcasas_c_Page_085.jpg
0eef2ce1f96528e688d2761adf341c66
14357d92ad783d3e58f66b4e386c84d967eba34c
93058 F20101114_AACWKU cruzcasas_c_Page_067.jpg
fa5efbc29fa08432e197ca24c1f2f7e4
998b9f306aa6ecacfef963afd6d64ce3826394d6
105155 F20101114_AACWLJ cruzcasas_c_Page_086.jpg
6153947d83c47a2f0b2e581f4497895c
eca1c0964445ee736f420ed232095fe1de7cfff8
115307 F20101114_AACWKV cruzcasas_c_Page_068.jpg
7f197c970edba167b9c3a8f860a526b6
320d054918de01b872c7cc23d182dc0cabac97f4
113625 F20101114_AACWLK cruzcasas_c_Page_088.jpg
80dfcd6f6e7efb8918f54a155d397c5b
8e012ce428fa154e2299b7cf2323dffde3d72e8d
111562 F20101114_AACWKW cruzcasas_c_Page_069.jpg
a4dc3c0dc8c140486e8b3aefe688acf2
f2a3cdb24bb38eea253f6d06f21c633e0f54dc80
25576 F20101114_AACWLL cruzcasas_c_Page_089.jpg
959946062786f809d73499c1db7dd7e9
cfa28ada047d3a0acdf397c0d86a9f45bb7ccc16
111486 F20101114_AACWKX cruzcasas_c_Page_070.jpg
715beb889501a4215caf481dbe523e0b
71f47037107ba1e2e9bc4e26ae63ffe1e9f9aa82
1051976 F20101114_AACWMA cruzcasas_c_Page_007.jp2
deeab7b087813d728361f5607c62cd57
7f8fb116557ea54b6e729f7ac82e98e4592c91b2
58780 F20101114_AACWLM cruzcasas_c_Page_092.jpg
5cfce43fa39db81c045ae983cd786842
32b4dbc47428afc05f4fb2801e97f12f0f94bf7b
98581 F20101114_AACWKY cruzcasas_c_Page_071.jpg
b3cafa40527634730e953b7498b726fe
298ea0a75a212cdd61d6b7f0115cc55231e1b57a