Citation
Development of a Flight Avionics System for an Autonomous Micro Air Vehicle

Material Information

Title:
Development of a Flight Avionics System for an Autonomous Micro Air Vehicle
Creator:
PLEW, JASON ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Accelerometers ( jstor )
Aircraft ( jstor )
Altitude ( jstor )
Global positioning systems ( jstor )
Internet search systems ( jstor )
Prisoners of war ( jstor )
Sensors ( jstor )
Servomotors ( jstor )
Transceivers ( jstor )
Vehicular flight ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Jason Plew. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
12/18/2004
Resource Identifier:
57731739 ( OCLC )

Downloads

This item has the following downloads:

plew_j ( .pdf )

plew_j_Page_58.txt

plew_j_Page_19.txt

plew_j_Page_80.txt

plew_j_Page_61.txt

plew_j_Page_83.txt

plew_j_Page_59.txt

plew_j_Page_04.txt

plew_j_Page_45.txt

plew_j_Page_72.txt

plew_j_Page_25.txt

plew_j_Page_56.txt

plew_j_Page_07.txt

plew_j_Page_75.txt

plew_j_Page_15.txt

plew_j_Page_86.txt

plew_j_Page_18.txt

plew_j_Page_42.txt

plew_j_Page_81.txt

plew_j_Page_73.txt

plew_j_Page_64.txt

plew_j_Page_82.txt

plew_j_Page_62.txt

plew_j_Page_50.txt

plew_j_Page_08.txt

plew_j_Page_47.txt

plew_j_Page_69.txt

plew_j_Page_12.txt

plew_j_Page_55.txt

plew_j_Page_06.txt

plew_j_Page_53.txt

plew_j_Page_63.txt

plew_j_Page_52.txt

plew_j_Page_85.txt

plew_j_Page_02.txt

plew_j_Page_77.txt

plew_j_Page_48.txt

plew_j_Page_46.txt

plew_j_Page_74.txt

plew_j_Page_03.txt

plew_j_Page_40.txt

plew_j_Page_33.txt

plew_j_Page_24.txt

plew_j_Page_44.txt

plew_j_Page_14.txt

plew_j_Page_76.txt

plew_j_Page_70.txt

plew_j_Page_88.txt

plew_j_Page_29.txt

plew_j_Page_05.txt

plew_j_Page_51.txt

plew_j_Page_38.txt

plew_j_Page_35.txt

plew_j_Page_01.txt

plew_j_Page_31.txt

plew_j_Page_16.txt

plew_j_Page_41.txt

plew_j_Page_23.txt

plew_j_Page_21.txt

plew_j_Page_79.txt

plew_j_Page_27.txt

plew_j_Page_84.txt

plew_j_Page_10.txt

plew_j_Page_32.txt

plew_j_Page_57.txt

plew_j_Page_13.txt

plew_j_Page_43.txt

plew_j_Page_39.txt

plew_j_Page_65.txt

plew_j_Page_67.txt

plew_j_Page_54.txt

plew_j_Page_30.txt

plew_j_Page_60.txt

plew_j_Page_66.txt

plew_j_Page_71.txt

plew_j_Page_11.txt

plew_j_pdf.txt

plew_j_Page_28.txt

plew_j_Page_26.txt

plew_j_Page_36.txt

plew_j_Page_68.txt

plew_j_Page_09.txt

plew_j_Page_17.txt

plew_j_Page_22.txt

plew_j_Page_49.txt

plew_j_Page_34.txt

plew_j_Page_37.txt

plew_j_Page_20.txt

plew_j_Page_87.txt

plew_j_Page_78.txt


Full Text












DEVELOPMENT OF A FLIGHT AVIONICS SYSTEM FOR AN AUTONOMOUS
MICRO AIR VEHICLE

















By

JASON PLEW


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

UNIVERSITY OF FLORIDA


2004


































Copyright 2004

by

Jason Plew

































This work is dedicated to my wife, Shalomoth Plew. Her love and support during this
research allowed me to not only accomplish my goals for this project, but also ensured
that my life was filled with happiness. It is also dedicated to my father, Richard Plew,
who made all of this possible. It was his support and guidance through the many years of
high school research projects that eventually led me to robotics, the Machine Intelligence
Lab, and eventually the MAV project.















ACKNOWLEDGMENTS

I would like to thank the professors affiliated with the Machine Intelligence Lab,

including Dr. Keith Doty, Dr. Antonio Arroyo, Dr. Eric Schwartz, Dr. Michael Nechyba,

Dr. Karl Gugel, and Dr. Michael Lynch, for providing me with the knowledge necessary

for this research. Their lessons both in the classroom and outside it were invaluable. I

would also like to thank the members of the AVCAAF research group. This of course

includes Dr. Pete Ifju and his students for developing the Micro Air Vehicles that we

would use for our platforms. Without their amazing vehicles, none of this would have

been possible. Thanks also go out to the controls team, led by Dr. Andrew Kurdila and

Dr. Rick Lind, who developed the navigational controller for the MAV128 R4. Among

their students, I would especially like to thank Mujahid Abdulrahim, for developing the

inertial-based controller for the MAV128 R5 to verify its ability to control a MAV, and

for sharing his knowledge and experience in RC aircraft and control systems. I would

also like to thank Dr. Nechyba and the other students in the MIL MAV group, and most

importantly Jason Grzywna for his development of the ground station system and

participation in this research.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ......... ................................................................................. iv

LIST OF TABLES ................ ......... .....................vii

LIST OF FIGURES ....................................................................... ......... ..................viii

ABSTRACT ............. ...................................................... ..............

CHAPTER

1 IN T R O D U C T IO N ...................................................................... 1

2 REVISION 3: PROTOTYPE FLIGHT SYSTEM ................................................... 5

Sy stem R equirem ents ....................................................................... 5
Flight System C om ponents .................................................... 7
Inertial Sensors: Microstrain 3DM-G .......................................................... 7
Global Positioning System Receiver: Axiom Swift A2 ..................................... 8
RF Transceiver: Microhard MHX-2400 ................................................... 9
The MAV128 R3 Onboard Computer........................... .................. 11
Flight System and Integration........................................................... ...16
Prototype Flight System Development: Conclusions................. ... ........... 24

3 REVISION 4: FLIGHT SYSTEM WITH ONBOARD GPS .................................. 26

Flight System C om ponents.......................... ...... ....... ............ ............. 27
Global Positioning System Receiver: Furuno GH-80D 2.............. ................ 27
RF Transceiver: Aerocom m AC4490-500................................. .................... 28
The MAV128 R4 and Flight System Integration............................................... 31
Revision 4 Flight System Conclusions................................................ .............. 44

4 REVISION 5: FLIGHT SYSTEM WITH ONBOARD IMU, GPS, AND
C O N TR O L L E R ..................................................................... ...... .............. 46

The MAV128 R5 Power System .............. .................................................. 48
Development of the Inertial and Analog Conversion Systems............................ 55
Flight Testing and Onboard Controller Development............................................ 63









5 CONCLUSIONS AND FUTURE WORK............ .......... ................. 69

L IST O F R E F E R E N C E S ............ .......... ...... .......... ................................................. 74

B IO G R A PH IC A L SK E T C H ............ .......... ............................................................... 77
















LIST OF TABLES


Table pge

3-1: Analysis of Thermal Dissipation Issues in Powering the R4 Flight System........... 37

3-2: Pressure Sensor Conversion Data... ...................... ................. 40

3-3: Flight System s W eight Distribution (Gram s)................................ .................... 44

4-1: Analysis of Thermal Dissipation Issues in Powering the MAV128R5, GH-80, and
A C4490 ............... ................................................. .... .... ......... 54

4-2: Conversion Form ulas for Inertial Sensors............................................ ... ............... 65
















LIST OF FIGURES

Figure page

2-1: 3D M -G IM U ........................................................................................... .7

2-2: Axiom GPS Radio.............................................................. 8

2-3: M H X -2400 RF Transceiver.................................................. ........................ 11

2-4: MAV128 R1, R2................................ 12

2 -5 : M A V 12 8 R 3 ............................................................... ................................... 13

2-6: M A V 128 R 3 D daughter B oards....................................................................... ... 15

2-7: Revision 3 Flight System Software Architecture ................................................. 19

2 -8 : 3 0 M A V .............. .... ............. ................. .............................................. 2 0

2-9: 3DM-G Benchtop Analysis .............. ................. .............. 22

2-10: 3DM-G Benchtop Analysis with Data Stabilization............................................ 23

2-11: M A V 128 Flight C ontroller......................................................... .............. 25

3 1 : G H -8 0 D .............. .... ............. ................. ................................................ 2 7

3-2: A C4490 RF Transceiver................................................................. .............. 29

3-3: Rev. 4 Flight System Software Architecture................................... 33

3-4: M A V 128 R4A R4B ............. ................................... ..................................... 34

3-5: SOT-223 Maximum Power Dissipation................................. ......................... 35

3-6: MAV128 R4C With AC4490........ ......................... .................. 37

3-7: A V CA AF M A V Platform 1.0 ......... ................. ............................ .............. 38

3-8: Revision 4 Flight Testing Setup .............. ............ .......... ................... 39

3-9: Rev. 4 Flight System Autonomous Waypoint Navigation................................... 43



viii









4-1: MAV128 R5A ....... ......... ............... ...........48

4-2: MAV128 R5A AC4490 Regulator ................... .............................. 49

4-3: TO-263 M aximum Power Dissipation .............. ................................................ 51

4-4: SO T-23 M axim um Pow er D issiaption................................................ .............. 51

4-5: MAV128 R5B................................ ....................... 52

4-6: MAV128 R5C Power System: First Stage and AC4490 Regulators..................... 53

4-7: M A V 128 Pow er System s ........................................................... .............. 54

4-8: MAV128 R5C................................ ....................... 57

4-9: M A V 128 R 5C Softw are A rchitecture............................................. ... ................ 61

4-10: MAV128 R5C Altitude Data .............. ............................ 62

4-11: A V CA A F 2.0 ...................................................................... ....... ..... 63

4 -12 : A ccelerom eter D ata............................................................................ .. ........ 67

5-1: Accelerometer Results, 6 Foot RC Aircraft Platform ........................................... 69















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 Science

DEVELOPMENT OF A FLIGHT AVIONICS SYSTEM FOR AN AUTONOMOUS
MICRO AIR VEHICLE

By

Jason Plew

December 2004

Chair: Antonio Arroyo
Major Department: Electrical and Computer Engineering

The goal of this thesis was to develop a complete avionics system for a Micro Air

Vehicle. This system must support the research UF is currently conducting in

demonstrating an autonomous MAV capable of operations in an urban environment. To

support this ability, an onboard system was needed that included both an Inertial

Measurement Unit (IMU) for stabilization of the MAV and GPS for navigational support.

There was also a need for an RF transceiver for linking the MAV to a Ground Station for

telemetry and control. A flight computer, designated the MAV128, was required to

provide an interface between all of these systems, and to eventually allow for an onboard

controller. Combined with an onboard camera, video transmitter, and vision-processing

algorithms on the ground station, this system was to allow us to begin developing a MAV

capable of the advanced maneuvers required for flying among buildings.

A prototype flight system was developed for a 24" MAV but could not demonstrate

autonomous control and navigation due to significant noise in the data from the IMU and









the overall weight of the onboard system. In the development of the next system, the

weight was reduced and onboard autonomous navigation was demonstrated, with flight

stabilization being handled by a ground-based horizon tracking system already

developed. We then integrated an IMU into the MAV128, and though development in

both a mechanical mounting system for the system and advanced control filters are

required to ensure valid inertial data, flight test results have been encouraging. Once this

version of the flight system has demonstrated autonomous flight, we will turn to

incorporating inertial, vision, and GPS control onboard, moving us even closer to our

goal of demonstrating autonomous urban operation of a MAV.














CHAPTER 1
INTRODUCTION

A major focus of the United States Air Force (USAF) in the next decade is the

development of Unmanned Air Vehicles (UAVs) that can be deployed in tactical

scenarios, as opposed to pure strategic operations, such as the Predator [1]. This includes

the ability for soldiers in the field to deploy Micro Air Vehicles (MAVs) for onsite

surveillance. The military also wishes to use these aircraft in complex environments, such

as urban areas, in multiple scenarios. Of course, as is the case with most military

technology, there are also numerous applications in the civilian sector.

The Micro Aerial Vehicle Lab at the University of Florida's Department of

Mechanical and Aerospace Engineering has been involved in the research and design of

such aircraft for several years, and has become very proficient in their development of

MAVs [2, 3]. Their technology in the development of these airplanes has led them to win

the International MAV competition for the past five years in a row because of their small

size and relatively long range. The airplanes of the MAV lab have traditionally been

controlled by off-the-shelf RC airplane equipment. To increase their range, the airplanes

often flew with a forward-looking camera and a video transmitter, allowing the pilot to

fly the aircraft when it was beyond the pilot's visual range. Given the size of the airplane,

the MAV pilots had to become quite skilled to keep the planes in the air.

The research of Scott Ettinger helped change the great skill requirement of the

pilots; he developed the first autonomous MAVs [4, 5]. In his research he took advantage

of the fact that many of the MAVs were sending a video signal to the ground, and









developed a system to analyze these images to find the horizon. A PID controller was

developed to take the position of the horizon in the image, and determine the necessary

commands to the servos to adjust the control surfaces to keep the horizon level. It could

also keep the horizon at a specific angle, allowing the aircraft to hold a specific roll angle

and thus orbit a position. The necessary servo commands were then sent through a device

that would convert them to signals understood by a standard RC controller, which would

then transmit the commands to the plane back to the airplane. Once it was working in real

time, the vision-based flight stability system could keep a MAV in the air without any

input from a human pilot. Further work led to the ability for the controller to take input

from ajoystick, allowing the MAV to be flown by any untrained pilot.

The success of Ettinger's work led to the initiation of a new project Active

Vision for Control of Agile Autonomous Flight (AVCAAF) [6]. The purpose of this

program, which began in 2003 and is to last for five years, is to develop Micro Air

Vehicles that are capable of autonomous flight within complex environments, such as

urban settings. However, there are multiple challenges that must be overcome to

accomplish this goal. Though autonomous flight had been demonstrated through the

horizon-tracking system, this alone was not sufficient for performing the complex

maneuvers required in any environment beyond an open field, such as a forest or city.

To be able to achieve the control necessary to fly a MAV autonomously in an

urban environment, we needed to tackle the problem at three levels. Traditionally, control

of aerial robots has been accomplished through an Inertial Measurement Unit (IMU).

Such devices, which often include accelerometers and gyroscopes, are very good at

determining the instantaneous movement of the airplane. Therefore, they can be very









accurate at tracking the movement of the vehicle over a short period of time. However,

over longer periods of time the error of the devices increases to the point that they are ill

suited to measuring the movement of the airplane. GPS has often been used to provide

the necessary data to recalibrate the IMU data. A controller can start with a known

position from the GPS, and track the vehicle's progress with inertial sensors, updating the

position to correct for errors on every GPS position update. However, the GPS unit can

only determine the position of the aircraft in relationship to the earth, and potentially

previously known fixed objects, if such data is available. It is unable to provide any real

information about the changing environment around the airplane. The vision system,

however, can be placed directly between the Inertial System and the GPS, as yet another

sensor to provide data to the airplane control system. A vision system's ability, for

example, to track a moving target or determine that an obstacle is ahead of the aircraft

allows it to provide information that would not be available to the more traditional

methods of airplane control. By using GPS, Vision, and an IMU, the inadequacies of any

one system are covered by the other two [7].

The first component of the vision system had already been developed, in the form

of the Horizon Tracking System. While the 900 MHz x86-based computer used in the

original tests were adequate for running the horizon tracking system, the transfer of the

horizon-tracking program to new computer technology ensured that enough processing

power was available for more complicated vision processing tasks. The technology for

GPS and IMU sensors that could be used for aircraft on the scale of even the largest of

MAVs was only in the initial levels of development. No off-the-shelf systems existed that

could be both flown on MAV, and integrated with the vision system already being









developed. Therefore, we had to develop the technology ourselves, i.e., develop a

complete flight system capable of demonstrating autonomous flights. We began

developing such a flight system in the spring of 2003.

While the Department of Defense defines a Micro Air Vehicle as an airplane with a

wingspan of less then six inches, the current level of technology was not advanced

enough to be used in such a platform. Therefore, we continued to use the 24-inch

wingspan systems used in the earlier vision-based flight stability experiments. Two initial

attempts to develop a flight system for a MAV were focused on identifying the proper

devices to use and to then develop a means of transferring telemetry from the airplane to

the ground. These systems also helped us to understand the problems we faced in

developing a MAV with the capability of performing autonomous urban operations,

which led us to the requirements for the third revision of the flight hardware. This third

system culminated in our first attempt at autonomous flight.














CHAPTER 2
REVISION 3: PROTOTYPE FLIGHT SYSTEM

System Requirements

The goals of the first year of the AVCAAF project were based on the realization

that the autonomous MAVs required Inertial and GPS capabilities in addition to vision.

The primary effort was to develop an onboard flight system for a two-foot wingspan

MAV that included both an IMU and a GPS receiver. This setup was developed to work

in conjunction with a new ground station including an enhanced vision-guidance system

based on Ettinger's work. The overall system requirement were to autonomously fly the

MAV through multiple GPS waypoints, and be ready for demonstration at Eglin AFB in

July 2003. Initially, the desire was to build as much possible functionality into the system

as possible. This included the ability to both send sensor data to the ground as well as

store it in the onboard system for later retrieval. Multiple locations of the controller (both

on the ground station and in the onboard system) were considered. We also wanted the

ability to control the servos through both the onboard flight system and through standard

RC equipment, allowing the computer to control the airplane and yet insure that a human

pilot could step in if necessary. As experimentation began, some of this functionality was

determined to be unnecessary. Other components proved to be inadequate for

accomplishing the task of developing a controller for the airplane, and were abandoned.

As a result, the prototype flight system that was developed by the end of the summer of

2003 had gone through several iterations, and the relationship between the different

components often changed.









There were multiple challenges in developing the onboard system. The major

issue was the weight constraint. A MAV with a wingspan of two feet, once outfitted with

motors, servos, batteries, and RC equipment, could only handle payloads of less then 100

grams. Furthermore, there was the difficulty in developing a system that was located

onboard the airplane access to the device when flying in the air was limited, and it

would be difficult to replicate these conditions in the lab. This was especially the case

when it came time to develop a controller to link the onboard sensors to the control

surfaces of the MAV.

Initially, it was hoped that a controller could be developed to function onboard the

aircraft. And as the research progressed, this remained the end goal of the flight system.

However, any time an onboard controller was in need of serious modification, either we

would have to land the airplane so that the onboard system could be reprogrammed, or

we would have to develop a robust method of transmitting a new controller to the

airplane. Since both of these options were considered to be unsuitable for the

development phase, we decided to develop the controller on a ground station. This

approach made it easier for the controls team to develop the necessary algorithms to fly

the airplane autonomously, allowing the fusion of the vision, inertial, and GPS data to

occur in a single location. The airplane needed to send all of its sensor data to the ground

station, and either process servo commands through standard RC equipment or through

the onboard flight system. Once the process of developing a controller for the MAV on

the ground had been refined, we began work on the development of transferring the

system to the onboard hardware.









Flight System Components

The major goal of the flight system was to have onboard inertial sensors and a GPS

receiver. Due to weight restrictions, very small sensors had to be found, even if the

tradeoff was in reduced accuracy. Furthermore, we needed to include the ability for the

flight system to transmit at least some of this sensor data to the ground. Finally, some

form of a processing system was required to tie all these individual components together,

and link them to the servo motors controlling the plane.

Inertial Sensors: Microstrain 3DM-G

Instead of attempting to build an IMU from discrete parts, we tried to obtain an

integrated component light in weight yet able to give us the accurate data necessary for

stable flight. The solution was the 3DM-G, an IMU from Microstrain (Figure 2-1). This

device included accelerometers in all three directions to give us the instantaneous

movement of the airplane, as well as gyroscopes for all three axes to provide roll, pitch

and yaw rates. Also included were magnetometers to give the orientation of the aircraft in

relationship to the Earth's magnetic field, and other assorted sensors. An onboard chip

would process the sensor data to provide not only filtered sensor data, but also the

orientation results (e.g., Euler Angles)

that are useful for controllers. The 3DM-

G could be communicated to through a

Universal Asynchronous Receiver

Transmitter (UART), a standard

communications device found on most 4


Figure 2-1: 3DM-G IMU









embedded processors, as well as on the serial ports of personal computers [8].

Global Positioning System Receiver: Axiom Swift A2

The other major system was the GPS unit. Once again weight was a major issue in

deciding on which device to select. However, in this case, the accuracy of the data was

not as important a factor. No GPS receiver currently available could return data accurate

enough to be useful in the flight stability problem, and therefore, the GPS was only used

for navigation. Because of this, a GPS unit that was only accurate to within several feet

was not an issue.

To save on weight, we looked for a GPS with an integrated antenna. Though using

an integrated antenna results in less capability in detecting the GPS satellites, especially

in extreme maneuvers, we decided that the savings in weight made such a choice worth

the degradation in navigational data. The resulting solution was the Axiom Swift A2 GPS

Receiver [9]. This device, shown with an interface in Figure 2-2, communicates both the

position and course of the aircraft through a UART, and had been used in previous work

that focused on exploring the possibility of integrating the horizon tracking system with

GPS [10]. Given the receiver's low weight

of 20 grams and the fact that we already had

some of these units available made this

choice for a GPS unit nearly ideal.

By default, the Swift A2

communicates using the industry standard

NMEA Protocol, which can provide

multiple navigational data, including Figure 2-2: Axiom GPS Radio
multiple navigational data, including









Longitude, Latitude, and Course Speed and Bearing. The drawback was that the NMEA

protocol uses ASCII to transfer data, which takes up much more bandwidth then using a

raw binary format. The onboard system first had to convert the GPS data into binary if it

was to be sent to the ground. While this was feasible, we wished to avoid having to waste

processing time parsing the ASCII stream. Fortunately, the Swift A2 could also be set to

communicate using the raw SiRF standard protocol. The interface to the GPS using this

standard proved to be much simpler. One minor issue was that the SiRF protocol, unlike

NMEA, communicated position data using a 3D axis system with the origin at the center

of the earth, the z-axis pointing to the north pole, and the x-axis along the equatorial

plane and perpendicular to the prime meridian [11]. For the development of navigational

controller to be straight forward, the format of the position needed to be in the latitude,

longitude, and altitude format. However, the conversion proved to be just a matter of

converting from the Cartesian coordinate system to a Spherical Polar system, a process

simple to carry out on the ground station computer.

RF Transceiver: Microhard MHX-2400

We needed some way of being able to communicate with the flight system when it

was in the air. The bandwidth could range between a few bytes for sending servo

commands up to the plane (bypassing the RC Radio System) and getting simple status

information back to receiving several bytes of telemetry including Inertial and GPS data.

The necessary data had to be sent at a rate of 30 Hz for the controller on the ground

station to function properly. We also had to be concerned with the range of the

transceiver. We needed to maintain the radio link for at least a few miles. One benefit of

the MAV system was that with one of the transceivers being in the air, we were operating

in close to Line of Sight (LOS) conditions.









Weight and size were as always a concern in determining the requirements for a

transceiver. We also needed to look for a device that operated in one of the unlicensed

bands of spectrum, so that we would not have to spend time getting approval from the

necessary government agencies to operate the system. These frequencies included 900

MHz, 2.5 GHz, and 5 GHz. The latter spectrum was only just beginning to be utilized,

and so the options for using a device at that frequency band were limited. We therefore

decided to use either a 900 VMHz or 2.5 GHz. Preliminary tests showed that using a data

transceiver and a video transmitter operating at the same frequency resulted in too much

video noise on the ground. We did have the ability to transmit video at either 2.5 GHz or

900 MHz, so it was an easy matter to keep the two devices from interfering with one

another.

We decided that it would take too long to develop our own transceiver, even using

the RF chipsets such as Intersil's Prism that have become popular in the past few years.

Therefore, an off the shelf solution was investigated. Initial work had been undertaken in

getting data to the ground from some of our initial flight systems using the CompactRF

Radio Transceiver from Microhard Corp. This device operates at 900 MHz, and can,

under optimal conditions, transfer data at up to 19.2Kbits for a range of 20 miles. [12]

Theoretically, we should then be able to send a packet of 50 bytes at a rate of about 20

Hz. However, even with the rather ideal conditions of having one of the radios in the air,

we were only able to send inertial data to the ground at around 10 Hz. Since this did not

meet our targeted data rate of 30 Hz, another radio needed to be found.









After an exhaustive search

for other off the shelf solutions, we

settled for another radio made by

Microhard. The MHX-2400

(Figure 2-3) advertised a 115.2K

data transfer rate at the same range

as the CompactRF radio. For that Figure 2-3: MHX-2400 RF Transceiver

matter, they offered the MHX-900

radio as well, which was identical to the MHX-2400 except that it operated at 900 MHz

[13]. Using the Microhard radios was attractive because it gave us flexibility in

determining which frequency bands the Datalink and Video systems would use.

However, a major drawback was the size and weight whereas the CompactRF was 2" x

1.5" at 20g, the MHX-2400 was 3.5" x 2" at 75g. However, there appeared to be no other

viable options at that time, and so this device was chosen to allow telemetry to be sent

from the MAV to a controller on the ground station.

The MAV128 R3 Onboard Computer

It was evident from the beginning that some sort of embedded processor was

required onboard the aircraft. There was initially some thought given to using an off the

shelf Power PC embedded system, running ucLinux. Such a device could be interfaced to

inertial sensors and a GPS, and it was powerful enough that it could easily run a

controller onboard, and possibly even part of the vision system. However, the size,

weight, and power requirements of this device led us to drop this as a solution, as did the

fact that its version of ucLinux required over a minute to boot. There was still the desire

to build a system that could handle the GPS and IMU systems for now, but eventually be









capable of supporting the horizon tracking

system onboard. Therefore, a DSP system

was considered. However, we only had a

few months to develop a flight system

capable of using the vision and inertial

sensors to keep the MAV stable, and using Figure 2-4: MAV128 R1, R2

GPS to navigate. Therefore, it was decided to just focus on the IMU and GPS for the

onboard systems. The vision processing systems would remain on the ground station

computer for now. For this situation, a normal embedded microcontroller could be used.

An Atmel AVR Mega128 microcontroller was selected, a device we had used in

past projects. This was the most powerful processor in the AVR family at the time,

capable of running at up to 16 MHz, and giving up to 16 MIPS of throughput. It contains

128K of Flash Memory for Programs, and 4K of RAM for data [14]. Besides its

performance and the fact that we were already familiar with this device, the Megal28

was well suited to the project for the following features. It had up to eight PWM outputs,

which were necessary for controlling servos, as well as an eight-channel 10-bit ADC for

reading analog sensors. While most microcontrollers have only one UART, the Megal28

has two, allowing control of two separate serial devices. A c compiler was available for

this device, allowing us to avoid assembly and therefore more quickly develop the code

for the flight system.









Development of an onboard flight

computer began in the spring of 2003. Due

to the fact that it was intended for small

aircraft and built around a Mega128, the

flight computer was designated as the

MAV128. The first two versions, shown in

Figure 2-4, were engineering prototypes and
Figure 2-5: MAV128 R3
targeted for general research into MAV

technologies as we determined the requirements of agile autonomous flight. The

MAV128 R3 was specifically targeted for the AVCAAF system demonstrated in July

(Figure 2-5).

We were unsure as to as to how well we would be able to isolate the flight system

from the motors, and keep the electrical noise they generated from interfering with the

system. Therefore, a ground and power plane were used with all of the flight computers.

Using these planes also allowed us to cut down on the number of traces that we needed to

layout on the design, and as a result minimize the size of the circuit board. We also saved

space by using small electrical trace widths and clearance requirements. Aside from

power signal traces, whose high currents required wider paths, all trace widths and

clearances were at most 8 mils (.008 in.), allowing us to keep the devices close together.

Even this specification proved to be too large, and we have since moved to 6 mils

clearance and trace widths.

Due to weight and size considerations, surface mount devices and other small

components were used whenever possible. While this was easy to accomplish in









procuring resistors, capacitors, and even most of the IC's, the matter of the connectors

was more difficult. To conserve space, we desired small micro-headers of .05" pitch

Some connectors had to remain at the standard header size of .10" pitch due to the fact

that the smaller connectors could not handle the current flowing through the cables. There

was also the possibility that external components such as the GPS, IMU, and RF

Transceiver would need to be connected in the field, and this was more difficult with

smaller headers. We were also unsure if they could handle the vibration of the plane.

Therefore, we continued to use standard headers for power and connections to all the

external components of the flight system (IMU, GPS, and transceiver). The ports on the

Megal28 itself used the micro-headers, significantly reducing the board size of the

MAV128. Small cables could then be made to connect to these headers, giving us access

to the internal peripherals of the Megal28. This method was used for the programming

port, but the cables were found to not be very secure, and therefore were not suitable for

flight.

A later development was that of daughter boards, which used the micro-header

ports to provide added functionality to the MAV128. Since multiple headers were used,

the daughter boards were far more secure then individual cables. Two such boards were

developed (Figure 2-6). The first was to allow the MAV128 to store the telemetry in

onboard external flash for later retrieval. The daughter board connected four separate

banks of 64 Mbit Atmel DataFlash to the SPI port of the Megal28. The major focus of

the project shifted towards getting telemetry to the ground for developing a controller,

and thus the data-logging capability was never used. The second daughter board allowed

the Megal28 to control up to four servos, and interfaced to their internal potentiometers









so that the MAV128 could record the

actual servo positions. This allowed us

to support the MAV used for the

research, which included three actual

servos as well as the drive motor,

whose speed was controlled in a

similar manner to a servo. The board
Figure 2-6: MAV128 R3 Daughter Boards
also included the Motorola

MPX4115A, a pressure sensor for measuring altitude, which could be read by the

Megal28's internal ADC.

In designing the MAV128, we also had to consider the power requirements of this

board. The principle power source for MAVs is currently Lithium Polymer batteries. The

large three cell batteries provide around 12 volts, while two cells provide 7.4 volts. The

major load on the battery, however, is the drive motor. The motor in our system requires

a three-cell battery. This results in a flight time of around 20 to 30 minutes. Since the

current draw of the onboard electronics is insignificant compared to the flight motor, we

did not have to worry about whether the electronics would be the critical factor in the

duration of a flight. Therefore, once again, weight was the primary factor. All of the

electronics on the MAV128 operated at 5V, and so the battery voltage had to be

regulated. If we were to use the same three-cell LiPoly battery as the motor, then the

difference between the input and output voltages would have been significant enough that

the thermal dissipation of the regulator could potentially result in the device overheating

and shutting down.









While the problem of thermal dissipation could have been completely eliminated

by using a Switching Regulator, such a device required external components, such as

large inductors and capacitors, and therefore increased the size of the board. We had to

keep the board as small as possible, and so instead we decided to power the electronics

on a separate two-cell battery instead, allowing us to use a standard voltage regulator in a

surface mount package, even though this resulted in a system with an efficiency of only

67%. Instead of a Linear Regulator, we used a LDO Voltage Regulator, which could

operate even when the battery input voltage dropped to less then 6V. This allowed us to

use the same electronics battery for multiple flights, and only have to replace the motor

battery to get the plane back in the air. The choice for the primary voltage regulator of the

system was a National Semiconductor's LM2940C. This LDO regulator could take the

battery input and produce up to 1A of current at 5V.

Flight System and Integration

In the end, all of the individual components had to be combined to work together,

and any issues ironed out so that a controller could successfully keep the airplane in the

air and navigate properly. A major focus was on the data and power interface between the

MAV128, 3DM-G IMU, Swift A2 GPS Radio, and the MHX-2400 RF Transceiver. A

major issue was the fact that all three devices connecting to the MAV128 required a

UART, but the Megal28 only had two UARTS available. Varying sources of power were

also an issue. The 3DM-G required at least 7V, as it uses its own regulators. The MHX-

2400 required 5V to operate, while the Swift A2 could only handle 3.3V. In the end, the

RF Transceiver was used on UART 0, and the 3DM-G on UART1. Code was written for

the Megal28 to allow one of the timer's output compare and input capture functions to

operate as a software UART. This pseudo communications port could only operate at low









speeds of 4800 bits per second or less, but this was not an issue for the GPS, which could

only update it's position every second. A standard connector was specified for all flight

hardware using a UART to make it easy to debug the system. Pin 1 contained data from

the MAV128, and Pin 2 data to the processor. Power and ground were designed to be on

pins 3 and 4, respectively.

We hacked the 3DM-G's cable so that it could connect to this standard UART port.

Interface boards were developed for the MHX-2400 and Swift A2 GPS so that they could

also interface to the port. The Swift A2 only required 150 mA at 3.3V. Therefore, 5V

from the MAV128 Regulator was sent through the UART2 port, and a LM3940IMP-3.3

voltage regulator from National Semiconductor was used on the interface board to

convert it to the GPS operating voltage. The LM3940 was chosen because it was in the

same package as the MAV128 voltage regulator, and was specifically meant for

converting 5V to 3.3V. The interface board also had the special connector necessary to

mate with the ribbon cable coming out of the Swift A2. As this cable was very small and

flat, we decided it was too difficult to hack, and left this cable in the system.

While it was felt that the MAV128 regulator could handle the current requirements

of the Swift A2 in addition to its own systems, this was not the case with the MHX-2400.

The RF Transceiver could consume more then 550 mA, over half of the current that could

be sourced by the LM2940. Therefore, another such regulator was used on the MHX-

2400 Interface Board. The MAV128 therefore sourced the raw battery current to the IMU

and RF Transceiver, and provide up to 1A at 5V to the processor systems and the GPS

Radio.









With the connections between the main components of the flight system finalized,

the code-base (written entirely in C) was developed to allow the Mega128 to interface to

both the GPS and IMU, and send this data to the ground in a binary format, as well as

receive commands from the ground station and adjust the servo values. The main

program was called datamav. A single global structure, called mav_data, was developed

that contained all of the data that was to be recorded from the sensors, and sent to the

ground station. This allowed all of the various software modules to access one location

for manipulating and transferring data, and make management of the code-base far easier.

We developed a standard interface for interfacing to the UART, which was then used for

all three communications ports in the system (including the software UART). The IMU,

GPS, and RF Transceiver software modules all went through this interface to transfer

data. Code was also written to read the servo commands from the RC system, and if

necessary, use these commands to control the servo motors of the airplane. However, this

function was not initially used for safety reasons, as it was considered necessary to

completely bypass the flight hardware at this stage of development. The end result was

the architecture seen in Figure 2-7.

After the flight system and code had been thoroughly tested in the lab, we began

flight tests. The MAV128, 3DM-G, and Swift A2 GPS were mounted into a 24" MAV.

As the effort to develop the ability to store telemetry onboard had been abandoned, the

MHX-2400 was also used to send the Inertial and GPS data to the ground. The total

weight of this system including the battery and Datalink antenna was 212 grams. The

flight system sent telemetry to the ground station through the RF transceiver. A controller

on the ground system was developed to use this data to send servo commands back to the




































SI









MAV128, which controlled the servos. While the controller was being developed, the

ground station passed through Servo Commands from a human pilot, allowing for open

looping testing

Even with open loop control, we experienced problems. Though some of this was

due to delays in controlling the aircraft through the ground station, another factor was the

over all weight of the flight system. It was at the very edge of the payload weight that the

24" MAV could handle, and as such, the airplane was extremely difficult to control.

Therefore, so that we could quickly start working on a controller, a 30" wingspan MAV

(see Figure 2-8) was developed for testing the prototype flight system. Once we had

successfully demonstrated autonomous flight, we could begin to focus on decreasing the

weight of the flight system.

It was at this point that the method of sending commands to the servos was also

changed, to eliminate the delays in controlling the airplane. The MAV128 just send data

to the ground, while the Controller on the Ground Station sent Servo Commands through

the RC System. This modification, along with using the 30" MAV platform, eliminated

the controllability problems, and

open loop testing began. Inertial

Telemetry and GPS data were both

sent back at about 30 Hz, which

was fast enough for the Controls

Group led by Professor Rick Lind ,

to begin work on developing a

controller. Figure 2-8: 30" MAV









More issues arose as open loop testing continued. The 3DM-G was set to output

both instantaneous acceleration rates and the Euler Angles that determine the orientation

of the plane. The data coming back from the IMU was incredibly noisy. It was initially

thought that earlier crashes that had resulted from the controllability issues had damaged

the 3DM-G, but when it was replaced with a second unit the same issues occurred.

Efforts at filtering were unsuccessful in cleaning up the data to the point that a controller

could keep the aircraft stably flying. There was a possibility that the mounting of the

3DM-G might be the issue. All of the flight system components were just placed inside

the fuselage, with foam keeping them from jostling around in the middle of a flight. We

thought that this might be too unstable for the 3DM-G, so wood inserts were used to

physically mount the 3DM-G to the airplane. However, there was still significant noise in

the 3DM-G data, such that the sensor was unusable.

This trend continued for the next few months as we approached the July 24th demo.

All efforts to develop a controller using the 3DM-G for inertial control proved

unsuccessful. Sources for this noise was thought to include EMF emissions from the data

and video antennas, or possibly from the drive motor of the MAV itself. The latter

seemed likely when bench tests showed perfect Euler Angle estimations from the 3DM-G

until the throttle was turned on. However, moving the 3DM-G away from these possible

sources did not help the matter.

Analysis by Jason Grzywna, Mujahid Abdulrahim, and myself found that a good

deal of the problem was the vibration of the airplane itself, and its affect on the 3DM-G

sensors. The MAV was put into a rig that allowed the throttle to be set to 100%, and yet

hold the airplane in place. The plots in Figure 2-9 show the 3DM-G sensor outputs both














i SO -
3.00 ~ ~ .,, --------------- ------------------------------------ --------------------------------------" -----~
o V A' lo-, "" zo^ ,i'" .- -- '-"^r^* ..l.s* ,'^I- _-1 ... __._.__ ._













r'o
-2, -








Figure 2-9: 3DM-G Benchtop Analysis


with the throttle active, and with the MAV undergoing vibration from external

movement, with the motor inactive. The vibration appears in both scenarios.

We tried to develop some form of advanced filter that would enable us to get useful

data out of the IMU, but with little success. However, a discovery was made that the

3DM-G did have the ability to produce filtered data. The IMU had the ability to use the

onboard gyroscopes to stabilize the data. As the plot in Figure 2-10 shows, doing so

immediately cleaned up the noise on the Euler Angles.

Even with this modification to the system, there were still issues with noise in the

inertial data. Another major issue discovered was that the accelerometers on the 3DM-G

could only handle gravitational force up to 2Gs. However, any extreme maneuvering










30,0












10Figure 2-10: 3DM-G Benchtop Analysis with Data Stabilization
PitLh
a0a03 1SG0 2 000 2500 30C.0 35 C" Oac-0 SOO sac--D Aileron


-200



-300


Figure 2-10: 3DM-G Benchtop Analysis with Data Stabilization


causes the aircraft to exceed these forces, and the 3DM-G accelerometer's to saturate.

These excessive forces were mostly happening during takeoff, but also occurred during

extreme maneuvers. While the instantaneous data would return to nominal values, the

process by which the 3DM-G calculates Euler Angles uses integration, and therefore each

successive case of high gravitational forces further distorted the IMU data. An attempt to

reset the calculation process both manually and at a set rate proved ineffective.

The process of experimenting with the 3DM-G was not helped by the conditions.

Multiple airplanes were destroyed during test flights, resulting in one 3DM-G being

destroyed. The rest of the hardware fared somewhat better, though the wear and tear

resulted in the small cable for the Swift A2 detaching from its connector inside of the

GPS radio. As a result, the metal case had to be peeled away (there was no way to open

the device) and wires soldered directly to pins on the internal circuit board of the GPS.

The MAV128, however, came away from all crashes unscathed.









Prototype Flight System Development: Conclusions

Our inability to get adequate data out of the 3DM-G made it impossible to develop

a controller by the time of the demo at Eglin AFB on July 24th 2004. This was the

primary reason that we were unable to demonstrate autonomous flight. Instead, parts of

the different MAV technologies currently being developed were shown, including a static

display of the prototype flight system.

There were other issues as well. The Swift A2 GPS had proved to be inadequate for

our requirements. Though it had the same weight as the 3DM-G, its large form factor

made it difficult to handle, especially since it had to be mounted on the airplane so that

it's integrated antenna could lock onto GPS satellites. Furthermore, the cable that it used

to connect to the other devices was too fragile, and extremely difficult to replace.

Besides the IMU problems, we also had to consider the overall weight of the flight

system. At almost 250 grams, it was simply too heavy for a 24" wingspan MAV to

handle. Though a great deal of this weight was the RF Transceiver and antenna, we still

needed to look into decreasing the weight of the other components as well. With aircraft

at this scale, every gram counted.

The one major benefit to come out of the Prototype Flight System was the

MAV128. The Flight Computer had proved to be more then adequate in handling the

multiple tasks placed on it, and had been durable enough to survive crashes in multiple

airplanes. As such, it was the one major component from this first system that continued

to be developed, with new versions remaining at the heart of our future flight systems.

Indeed, the MAV128 R3 has since been used in other MAV projects at the University of

Florida. An R3 board was used in the ground station to provide an interface between the

Ground Station Computer and an RC Controller. It also has been used for servo control









and transmitting telemetry in larger

aircraft for flight tests supporting

",Aiiie MAV research. The technology

was even moved into the

PocketMAV, a small 12" wingspan

MAV that is to allow for

augmented control of a Micro
Figure 2-11: tMAV128 Flight Controller Aerial Vehicle. The initial system
Aerial Vehicle. The initial system

actually used a MAV128 R3 with the addition of a daughter board containing a two-axis

accelerometer, and the integration of the two PCBs and the removal of unnecessary

components for the system led to the tMAV128 (Figure 2-11), a circuit board with

dimensions .7" x .7" [15].














CHAPTER 3
REVISION 4: FLIGHT SYSTEM WITH ONBOARD GPS

With the July demo complete, attention turned to the next demonstration, which

was supposed to occur in late October. We still had the goal of developing a flight system

to enable autonomous flight and navigation of a 24" MAV. To accomplish this, we

needed to both replace the 3DM-G and dramatically reduce the weight of the flight

system.

It appeared that the only way to create a controller capable of stabilizing the aircraft

during flight was to develop our own IMU system, directly integrated with the MAV128.

However, doing so in less then three months was near impossible, especially since we

also needed to also find replacements for the RF Transceiver and the GPS to cut the

weight of the overall flight system, and develop the Revision 4 MAV128 to support the

new system requirements. A temporary solution was to use the vision-guided stability

system. Instead of having an IMU onboard for the October demo, the controller used an

augmented version of the horizon tracking system to obtain the instantaneous roll and

pitch angles. This data could then be used to keep the aircraft stable while flying. The

new onboard flight system used the MAV128 R4 to obtain GPS data and send it to the

ground for the controller to use in navigating the plane. Furthermore, the ground station

also was used to develop new vision algorithms for objectives such as target testing. This

allowed the Revision 4 system to be used as a testbed for both advanced vision

processing systems and GPS-based navigational control [16, 17]. Once the October demo

was complete, we could then start working on developing on the MAV128 R5, which









would include Inertial Sensors and be a drop in replacement for the new system. The

Revision 5 board would allow the controller to be modified to take advantage of GPS,

Vision, and INS, and fully demonstrate autonomous flight.

Flight System Components

Global Positioning System Receiver: Furuno GH-80D

There were multiple reasons for replacing the Axiom Swift A2 GPS Receiver. A

major issue was it's weight, which was 28 grams. A related problem was the form factor.

At 1.65" by 1.65", the Swift A2 took up a major portion of the top surface area of the

MAV. Finally, it's connector had been proven to be unsuitable for the sometimes rough

conditions of MAV flight testing, with the result being that there were several trips to the

field where a lose connection resulted in no navigational data.

A search for a new GPS unit was begun. Another MAV grant at the University of

Florida had come across a new GPS unit, the GH-80 (Figure 3-1) by a company called

Furuno, and was having some success with it. After some investigation, we settled on the

GH-80 for the Revision 4 Flight System. At .8" x .8" size, and with a weight of only 12

grams, its physical specifications were well

suited for our needs. Furthermore, there was

a version available with 10 small pins

protruding from the bottom of GH-80 that

could be used for interfacing [18]. A simple

circuit was designed that allowed the GPS to
Figure 3-1: GH-80D
be directly soldered to a PCB, and then Figure 3-1: GH-80D

connected to the MEGA128 through the standard 4 Pin UART Header. This completely

eliminated the connector problems we had been experiencing with the Swift A2.









The protocol used to interface to the GH-80 was a manufacturer-defined binary

specification, but was similar enough to SiRF that it was quite easy to rewrite the GPS

interface code from the Swift A2 to work with the Furuno GPS instead. One issue that

was discovered was that the GH-80D ignored any commands sent to it for the first few

seconds after power up. Unfortunately, this information was not included in the

preliminary datasheet, and as such it took awhile to determine why the GPS units were

ignoring the commands being sent that set which data format the device should send.

Once this issue was resolved by continuously sending the commands until there was a

change in what data was g sent, the code necessary to interface the GH-80D to a

MAV128 flight computer was complete.

RF Transceiver: Aerocomm AC4490-500

The Microhard MHX-2500 had been selected primarily because of their advertised

ability to transmit at up to 115.2 kbps at 20 miles. Though they operated adequately, the

excessive weight was a critical issue. However, after an exhaustive search nothing else

could be found that met the requirements, and so the Microhard Transceivers were

chosen for the Revision 3 Flight System in April 2003.

It turns out that we should have continued looking for viable alternatives. One

month later, a company called Aerocomm started publicizing its new 900 MHz RF

Transceiver, the AC4490 (Figure 3-2), which could also communicate at 115.2K. The

device came in a few different versions capable of transmitting at different power levels.

The AC4490-500, which was the most powerful, had a range of 20 miles, more then

adequate for our needs. Furthermore, the device was 1.9" x 1.65" at a weight of only 12

grams, and therefore less then half the size and 1/7 of the weight of the MHX-2400 [19].









There were of course issues. One was the fact that the AC4490 was a very new

device. Fully detailed specifications on the device were not yet available, and so some of

our work in making the transceivers function in our system was trial and error, since the

necessary information was not always at hand. Furthermore, the AC4490-500 was not yet

available. While it was initially supposed to be ready in the summer of 2003, it ended up

being delayed and we were not able to

get the first two modules until late / *

** *
August. As a result, we initially started o

using AC4490-200 RF Transceivers *

instead. They were identical to the *

AC4490-500 modules except that they

transmit at a lower power, which

results in a maximum range of four
Figure 3-2: AC4490 RF Transceiver
miles. For initial testing, this was not

an issue.

We started with development kits for the transceivers. However, we ran into some

problems in testing the devices, which did not operate very well in the lab. We assumed

that the issues when testing the radios under these conditions was the result of

interference from other devices in the lab also running at 900 MHz, as well as from the

fact that we were working in confined rooms with walls that probably contained some

metal shielding. As a result, radio reflections were likely to occur. In any event, we found

we had better results when we kept the transceivers in separate rooms, and much better

performance once we moved outdoors.









There was the matter of determining how to set up the modules. Fortunately,

Aerocomm preset most of the necessary parameters for optimum performance depending

on the scenario the AC4490's were operating under. There was still the issue of setting up

the network. First of all, one AC4490 had to be set as a server, and the other as the client.

We also had to determine whether the server radio should be on the MAV, or stay with

the ground station. Furthermore, we needed to determine whether the master radio should

just communicate to the client AC4490, or broadcast to the world. If we chose the former,

every time we changed clients, we would have to modify the configuration of the server

radio with the new client address.

There was also the matter of whether the radios should be set to stream mode, or

acknowledge. Stream mode could actually result in a faster throughput, but as a result,

data packets were sometimes broken up. This latency in the received data appeared to

cause issues in remaining in synch with the ground station, and result in corrupted

telemetry. Furthermore, the radios would not retry sending corrupted packets in this

mode, and so errors could occur [20]. Therefore, acknowledge mode was selected. We

also adjusted the baud rate down to 19.2K, as sending so much data at higher baud rates

appeared to result in dropped packets as well.

Broadcast mode also seemed to result in dropped packets, and so we set the server

radio to communicate with the single client. We also noticed issues where the client radio

occasionally lost synch with the server. While the device usually reacquired the synch, at

other times the client needed to be reset. As this was impossible to do with the MAV

AC4490 when it was in the air, we chose to locate the client device in the ground station.









We had begun to gain an understanding of the devices when we received our first

AC4490-500s. However, we then ran into problems because of the fact that the higher

power radios were not compatible with our development boards. The AC4490-200 was

available in both 5V and 3.3V versions. We were using the 5V versions to have less of a

voltage drop when converting from the battery voltage to the RF Transceiver voltage.

However, it was then announced that the AC4490-500 would only be available at 3.3V.

The development boards Aerocomm had originally sent us were only an initial design

that could not support 3.3V modules. Fortunately, we were able to obtain new

development boards from Aerocomm that had much greater functionality, including the

ability to interface to our new RF transceivers. We tested the range of the AC4490-500s

and found that they were roughly equivalent to the old MHX-2400s. Satisfied, we now

focused on the development of the MAV128 R4, and the integration of all of these

components into the flight system necessary for the October demo.

The MAV128 R4 and Flight System Integration

The MAV128 R4, whose design began in early July 2003, grew out of a desire to

combine into a single design the MAV128 R3 and the daughter board that added servo

motor support and an altimeter, as well as integrating any necessary interface boards for

the new GPS and RF transceiver devices. Since the first two prototypes (there were three

different versions of the MAV128 for this version of the flight system R4A, 4B, and

4C) were created before it had been finally decided to drop the 3DM-G, both still

supported the IMU. The final version of the MAV128 R4 removed 3DM-G support.

Thereafter the RF transceiver was located on UARTO, and the GPS on UART1. This had

the added benefit of removing the need for the software-based UART2, and as a result the

output compare and input capture timing resources formerly reserved for the UART









became available for other functions. The latter was especially important, since it

reopened the possibility of reading servo commands from the onboard RC equipment.

The daughter board had been created specifically so that the servos would not have

to connect to the micro headers. Therefore, the servo ports on the R4 also used standard

.1" pitch headers. Servo feedback was once again available using the Megal28's 10-bit

ADC. The converter also supported the MPX4115A pressure sensor for measuring

altitude, which was also integrated into the R4.

The program that we developed for the MAV128 R4 was actually rather similar to

that of the Revision 3 system. The 3DM-G code was of course removed, and the GPS

interface software modified to process the Furuno Packets instead of SiRF. With the

removal of the software UART, the Megal28's input capture 3 hardware was available,

and code that had been developed earlier to read servo commands from an RC aircraft

receiver was reinserted into the program. In an attempt to eliminate synchronization

problems that sometimes occurred between the MAV and the ground computer, the

MAV128 only sent a packet of telemetry when it received a request from the Ground

Station. However, the software mostly remained the same, as seen in Figure 3-3.

Most of the changes through the series of MAV128 Boards resulted from

determining how to supply the AC4490 with power. Much of these issues were due to the

fact that we had access only to preliminary data. The AC4490 included both 5V and 3.3V

versions. We therefore designed the first MAV128 R4 board to supply 5V from the main

voltage regulator onboard (Figure 3-4). A connector located on the bottom of the board

allowed the AC4490 to connect directly underneath the MAV128, and communicate

directly to the ATmegal28.







33















f0
el





[ 0


Ig E



0
*S a

II i 3$ i
E~ E I ? ~ l





















~ ti
!. M l
E E l. ^ I




I~~ B j-- *




s ,W "









This procedure worked

rather well for the AC4490-200,

and the onboard supply was able to

power the AC4490, the MAV128,

and the GH-80, whose 3.3V supply

was derived from the same source.

However, it was then announced

that only 3.3V version of the
Figure 3-4: MAV128 R4A, R4B
AC4490-500 were available. Given

the fact that no power consumption data was available, we assumed that transmitting .5W

at 3.3V consumed about 150 mA. We had already selected a voltage regulator for the

GPS to replace the old LM3940, and decided to use it for the AC4490-500 as well.

Selected for its very small size of 3 x 3 mm, the REG113NA-3.3 could source up to 400

mA. As the GPS itself could consume almost 90 mA, we thought it better to use separate

regulators for the two 3.3V devices on the MAV128 R4B (Figure 3-4).

However, we immediately had problems transmitting data through the AC4490-500

when using the MAV128 R4B to source the power. It only worked for a minute or two,

and then only with intermittent communication. The In Range Signal on the Client

AC4490, which should assert when it is synched with a Server, constantly switched on

and off. We immediately suspected that the device was actually consuming more power

then originally thought. An interface board was created using the old LM2940 / LM3940

system from the Revision 3 System to convert a battery source to 5V and then to 3.3V.

However this system had the same problems.










Our attention then turned to 5

4 C S01-22$ PI CK G
whether placing the board PCB UN
I s, I 4 COPPER
3
underneath the MAV128 could be 2 OUNCE COPPER
2 2
resulting in some type of electrical
1 OUNCE COPPER
noise interfering with the operation 0 -
-40 -25 25 75 125
AMBIENT ItMPERATrRE (oC)
of the AC4490. We occasionally saw o0soa1

indications that this could be a Figure 3-5: SOT-223 Maximum Power Dissipation1

factor, such as moving the

transceiver away from the MAV128 resulting in the device transmitting again. However,

in most cases the position of the device appeared to have no affect on whether the RF

transceiver was operating, and we decided that any indications of interference from the

MAV128 being a factor was a coincidence. We were also unsure if maybe some signals

that were dealt with on the development board, such as RTS and CTS, needed to be

interfaced on the MAV128 Board as well for the device to operate properly. However,

the AC4490-200 had not demonstrated any problem in this area, and the datasheet said

that they could be left floating if unused. To verify, we made sure all of the inputs to the

AC4490 other then TXD, transmit data, were deasserted. Doing so had no affect.

Eventually, feedback from Aerocomm as to the actual power consumption of the

device determined that the problem had all along been providing enough power to the

AC4490-500. When transmitting 100% of the time, the radio consumed up to almost .5A.

Though this explained why the RF Transceiver was not working when powered by the

MAV128 R4B, it failed to explain why the interface board did not work.


1 LM2940 Datasheet, National Semiconductor, 2003









We soon determined the cause however we had failed to ensure the conditions

used for the LM2940 / LM3940 Power System on the interface board were the same as

the LM2940 / REG113 system on the MAV128 R4B. We had always powered the flight

computer on a two-cell battery, with a source of around 7.4V. When we powered the

interface board, however, we had used a bench-top power supply sourcing around 12V.

While the LM2940 can handle an input voltage of over 20V, there is also the matter of

being able to dissipate the thermal energy that is generated by the differential voltage

between the input and output multiplied by the current being sourced. In this case, the

LM2940 needed to dissipate about 3.5W. As seen in Figure 3-5, this was beyond the

capabilities of the SOT-223 surface mount package we were using, especially since the

device is meant to dissipate heat through the ground plane of the PCB. We were using a

prototyping device to quickly create the interface boards, instead of having them

fabricated and sent to us after a weeks delay. The drawback was that the circuit boards

used only one or two layers, with no internal planes. As a result, the voltage regulator had

little ability to dissipate thermal energy, and the load we were placing on it caused the

device to shutdown to protect itself, and only turn back on when its internal temperature

had returned to normal levels. Once we used a source of 7V for the AC4490-500

interface board, it worked properly.

With the proper information and the analysis of the power issues shown in Table 3-

1, we determined the requirements of the AC4490, and what devices were required to

provide power. Therefore, the LM2940 / LM3940 system was also used in the MAV128

R4C, the final version of the flight computer intended for the Revision 4 flight system

(Figure 3-6). As always, a 2-Cell LiPoly battery was sued to power the electronics, and as









a result the AC4490-500 was

able to transmit without the

power system going into

thermal shutdown. Later flight

tests verified that the RF

Transceiver could send data to

the ground, though Figure 3-6: MAV128 R4C With AC4490

transmission speeds when

sending about 50 bytes of data were limited to around 10 to 15Hz.

Table 3-1: Analysis of Thermal Dissipation Issues in Powering the R4 Flight System
Voltage Regulator Thermal Dissipation PT = I (VOUT VIN)
AC4490-200 (5V) Current Consumption 106 mA
AC4490-500 (3.3V) Current Consumption 492 mA
MAV128R4 (5V) Current Consumption 10 mA
GH-80 (3.3V) Current Consumption 88 mA
Two-Cell Battery, MAV128R4A, AC4490-200, GH-80
LM2940 5.0V I(7.4V 5.0V) (106 mA + 10 mA + 88 mA) = .4896W
GH-80 REG113NA 3.3V (5.0V 3.3V) 88 mA= .1496W
Two-Cell Battery, MAV128R4B, AC4490-500, GH-80
LM2940 5.0V | (7.4V 5.0V) (492 mA + 10 mA + 88 mA) = 1.416W
AC4490 REG113NA 3.3V (5.0V 3.3V) 492 mA= .8364W
GPS REG113NA 3.3V (5.0V 3.3V) 88 mA= .1496W
12V Power Supply, AC4490-500 Interface Board
LM2940 5.0V (12V 5.0V) 492 mA = 3.44W
LM3940 3.3V (5.0V 3.3V) 492 mA = .8364W
Two-Cell Battery, MAV128R4C, AC4490-500, GH-80
LM2940 5.0V | (7.4V 5.0V) (492 mA + 10 mA + 88 mA) = 1.416W
AC4490 LM3940 3.3V (5.0V 3.3V) 492 mA= .8364W
GH-80 REG113NA 3.3V (5.0V 3.3V) 88 mA= .1496W


With the power issues concerning the AC4490 complete, our attention turned to

conducting flight tests of the integrated flight system, which included the MAV128 R4C,

the Aerocomm AC4490-500, and the Furuno GH-80D GPS. Direct control of the servos









through the MAV128 was possible in two ways, with commands either sent from the

Ground Station through the AC4490 datalink, or by reading the servo commands from

the onboard RC Receiver through the Input Capture on the Megal28. However, we

instead used the setup shown in Figure 3-8, with the servos remaining on the standard RC

control system, and the ground station flying the MAV through the RC Controller.

Therefore, all the MAV128 R4C had to do was read in GPS Data, as well as take

readings from the altimeter through the ADC. All of this data was then sent to the ground

station through the datalink, where the controller combined it with the pitch and roll

angle estimations from the horizon tracking system to fly the plane.

Given the fact that we had significantly reduced the weight of the flight system, it

was decided to target a 24" platform again. A MAV that had been specifically designed

for the AVCAAF project (Figure 3-7) was therefore used in our flight tests. We did

notice some issues with the GH-80 GPS receiver once we began flying. At times, it took

up to a few minutes to lock on to enough satellites to be able to determine a position. It

was obvious that the issue was entirely with the GPS set, and not the MAV128 or

AC4490, because of the fact that we were

getting telemetry on the ground at all

times in these situations. Furthermore, the

timer on the GPS was being retrieved and

sent to the ground, and it was constantly

increasing, indicating that the GPS was

operational and communicating with the

MAV128. Figure 3-7: AVCAAF MAV Platform 1.0













ELECTRONICS SENSORS
-Atmel MAV128 -Camera
-Data Transceiver -GPS
-Video Transmitter -Altimeter
-Antennas
-Batteries

On-Board MAV
Computer








Ground Station
Tranceivers
(900Mhz 11Skps)


On-Board Camera



Video Recevier /
Sony Walkman


Hi-Tech Device Fulaa Radio
S(Signal Generation/ 4-9 io
Futaba Feedback) Coroller


Figure 3-8: Revision 4 Flight Testing Setup



Since the time it took for the GH-80 to start providing accurate position data varied,

we thought it was possible that the issue was the locations of the GPS satellites in orbit,

and that at certain times they were positioned in such a matter to make it more difficult

for the device to obtain a lock. Communication with Eglin AFB, where the device was

also being used, included observations that the GH-80D needed a sizable ground plane

beneath it to improve the performance of it's integrated antenna. However, a Furuno

Field Application Engineer claimed that such a modification was not necessary, and

though we did add copper plating to the hatch where the GPS Receiver was mounted, no

real improvement was detected. In any event, it often took no more then a few minutes

for the GPS to acquire a lock, and so the matter was dropped. As a precaution, the copper

plating remained on the hatch for all of the Revision 4 flight system testing.









The other issue that was observed as flight testing continued was the sensitivity of

the altitude sensor. This device registers the difference in pressure between vacuum and

the outside environment, and could range from 15 to 115 kilopascals (kPa). However, this

translated to being able to register an altitude in the range of several thousand feet,

whereas we were only concerned with registering changes on the order of a few hundred

feet at most. It was thought that the 10-bit ADC of the Mega 128 was capable of giving

the system the necessary resolution of a change in a few feet so that the MAV could hold

an altitude. However, initial results indicated a resolution of around 40 feet.

Given the fact that the Megal28's 10-bit ADC had a resolution of about 5mV when

using a 5V reference, we had the ability to register a change of about .1 kPa in pressure.

However, this only translated to an accuracy of around 30 feet, as shown in Table 3-2.

Furthermore, this was assuming that the ADC in the Megal28 performed with absolute

accuracy. In actuality, noise from the processor core could induce errors in the results,

and as such, our ability to read changes in altitude was reduced even further unless we

were to shut down the core while taking analog readings. Obviously, this was not the

solution.

Table 3-2: Pressure Sensor Conversion Data
ADC Sensitivity (Vrefgh VrefL 1 .0
Vrefhgh -Vrefo* 2I = .00488V
Pressure Sensor Sensitivity 45.9 mV / kPa
ADC Pressure Sensitivity 45.9mV 5mV
APS = .1089kPA
(APS) lkPa APS
Altitude vs. Pressure Increase of 1 foot = Decrease of .003559 kPa
ADC Altitude Sensitivity Ift APS
--> APS= 30.598feet
(AAS) .003559kPa .1089kPA


Instead, it was decided to modify the altimeter circuit. The MPX4115A was

initially soldered directly to the board. One was removed from a MAV128 R4C, and a









header installed that could connect to a prototype board whose purpose was to increase

the sensitivity of the altimeter system.

A few different methods were tried. One of the first attempts was to use operational

amplifiers to multiply the signal. Before it was multiplied, a set bias voltage was

subtracted from the sensor output so that the pressure range of interest (sea level to at

least a few hundred feet) existed between 0 and 5V. The trouble was that as the gain

factor increased, range of pressure detectable by the system decreased. The pressure at

ground level could vary depending on location and weather conditions, meaning that if

we set the gain factor too large, we risked being unable to detect changes in altitude

under certain conditions. Though several gain factors were considered and used during

experiments, a factor of eight was used in all flight testing, since it was the minimum

factor to produce a theoretical resolution of 4 feet, and we were concerned that a further

increase in gain would result in the ADC saturating.

To ensure that the system was simple, and that we could achieve the maximum

range possible, we required small single supply operation amplifiers that could produce a

true rail-to-rail output. Though such devices were difficult to find, we did have success

with National's LM324 and Linear Technology's LT1006. There was also an attempt to

bypass the Megal28 ADC and use a TI TLC3545 14-Bit Serial ADC, which then

communicated with the processor through a SPI Port. It was hoped that the fact that this

device was isolated from the Processor Core combined with its greater accuracy would

result in better altitude data. However, using the TLC3545 did not provide any

observable increase in accuracy over the original operation amplifier test system, even

when the altimeter input signal was preamplified by the op-amps.









We therefore retuned to working with just the operational amplifier system of

subtracting the bias voltage from the sensor output and then multiplying the result by

eight. The initial design used a single operational amplifier to keep the size of the circuit

down. However, we found that this approach allowed noise on the bias to have a major

affect on the output. As we were using a resistor divider network to generate the bias, this

was the signal most susceptible to noise. Therefore, we began using other op-amps to

buffer the bias voltage, the sensor output, and eventually even the analog input to the

Megal28 ADC, all in attempt to isolate the noise from the sensor system. However, we

still had at most a performance of around 20 feet.

Neither amplifying the circuit by a factor of 8 through the use of the operational

amplifiers, nor increasing the accuracy of the ADC by a factor of 16 (through the

TLC3545), had been adequate in increasing the performance of the Altitude Sensor. Even

combining the two systems had had no discernible affect. It was decided that if we were

to assume that the sensor was capable of providing the resolution necessary, then we

needed to further protect the altitude sensor from possible interference. The next step was

to isolate the power for the Mega128 from the sensors. One 5V voltage regulator had to

be the source for digital power and ground, another for analog. However, such a power

system required a complete redesign of the MAV128. Therefore, it was decided to not

spend time making this change for the Revision 4 Flight System, and instead this was

incorporated into the many changes necessary to turn the MAV128 into a complete IMU

system for Revision 5.






43



31 25


-6
S31 _















longitude min from 82 W (min)


Figure 3-9: Rev. 4 Flight System Autonomous Waypoint Navigation



While the experiments on the altitude sensor system were being conducted with

one MAV128 R4C, the other was being used in flight tests for developing the controller.

Initial attempts to use the ground station based horizon tracking system to keep the

aircraft stable were successful [21]. The controls team then moved on to creating the

outer loop of the controller so that the GPS data could be used for waypoint navigation,

or flying between different set points [22]. As seen in Figure 3-9, this controller allowed

the MAV to continuously fly through a series of GPS waypoints, and therefore

demonstrated our first autonomous MAV capable of performing missions.









Revision 4 Flight System Conclusions

We were able to develop a flight system that, when used with the pitch and roll

estimated derived from the ground, demonstrated autonomous control of the MAV. All of

the major systems used the MAV128 R4C, the Furuno GH-80D GPS, and the

Aerocomm AC4490 RF Transceiver had proved to be more then adequate for the job.

We had also successfully cut the weight of the overall flight system by more then half-

from 212 grams in July 2003 to 86 grams by October (Table 3-3). Furthermore, the

hardware was a major step towards developing the Revision 5 system with an onboard

IMU.

Table 3-3: Flight Systems Weight Distribution (Grams)
System Prototype MAV128 / GPS MAV128 /INS / GPS
Flight Computer MAV128 R3 16 MAV128 R4 10 MAV128 R5 16
INS 3DM-G 30
GPS Swift A2 28 GH-80D 12 GH-80D 12
Datalink MHX-2400 86 AC4490-500 12 AC4490-1000 12
Datalink Antenna 26 26 26
Battery Two Cell Li-Poly 52 Two Cell Li-Poly 52 Motor Battery -
Total Total 238 Total 112 66


One major issue that still remained was the problems we were having with the

MPX4115A altitude sensor. We were never able to reliably increase its sensitivity, even

when using the amplification circuits. This was a problem that had to be addressed as we

moved on to development of Rev 5. For one thing, we needed to determine a way to

detect small changes in the altitude of the MAV on the order of a few feet to be able to

develop a robust controller. It appeared that when powered on the same supply as the

digital components, the 10-b ADC on the Mega128 was not adequate for reading the

sensor output of the MPX4115A. This meant that we might also run into issues as we

integrated analog inertial sensors into the system as well. Therefore, the analog system of






45


the MAV128 was a major issue that would have to be addressed as we moved into the

development of the Revision 5 Flight System.














CHAPTER 4
REVISION 5: FLIGHT SYSTEM WITH ONBOARD IMU, GPS, AND CONTROLLER

The Revision 4 system had been debugged and was being used in test flights, and

development of an autonomous controller utilizing the vision system and GPS was now

underway. We therefore turned our attention to developing the Revision 5 flight system.

The overall architecture of the system remained the same, with the AC4490-500 being

used for the Datalink, and the GH-80D providing navigational data. However, the

MAV128 received a major upgrade, with the necessary changes made to support an IMU

system onboard as well as continue with all of its other functions.

The issues we faced in developing the MAV128 R5 actually were rather similar to

those faced in developing the Revision 4 system. And as before, it took three versions of

the MAV128 R5 before we got all of the major problems sorted out. We had been asked

to further decrease the weight of the overall flight system. The only solution was to no

longer use a separate battery to power the electronics, but instead run off of the same

battery that powered the MAV motors. The major issue with this was that the input

voltage to all regulators handling the battery supply was now 12V instead of 7.4. This

resulted in much higher thermal dissipation requirements then had been experienced with

the two-cell LiPoly batteries.

We also wanted to ensure that the MAV128 R5 remained the same size as the R4C.

However, we were adding a large number of new components to support the IMU.

Therefore, some changes to the overall layout were made. Whereas previous flight

computers used the ATMEGA128-16-AI, which came in a Thin Quad Flat Pack (TQFP)









package, we started using the ATMEGA128-16-MI with the Revision 5 computers. This

device came in a Micro Lead Frame (MLF) package 1/3 the size of the QFP. To further

conserve board space, the micro headers providing access to all of the individual ports of

the MEGA128 were removed. We felt that they no longer had any real purpose, due to

the fact that we had not really used them after abandoning the daughter board concept

after the MAV128 R3. Since then, redesigns of the boards had been used to incorporate

new functionality into the MAV128. Furthermore, there was also the fact that many of

the port functions were already being utilized by the MAV128 hardware. The number of

available peripherals inside the processor was rapidly decreasing, and so there was never

any real reason to connect a device to most of the microheaders. We also eventually

began to use the JST ZH/ZR connector system for the servo cables. Using these ports on

the MAV128 saved a lot of room due the fact that they were half the size of the old

connectors, though the drawback was that any MAV that was to be flown by the

MAV128 had to have it's servos modified to use the JST connectors instead of the

standard 0.1" pitch headers.

Even after we handled all of the problems with these transitions, there was still the

matter of developing an analog processing system for the inertial sensors that was

accurate enough for the controller. During the development of the Revision 4 flight

computer, we had immense difficulty in creating a processing system just for the

MPX4115A altitude sensor, and were not able to obtain the sensitivity that was needed.

Now, instead of just one sensor to deal with, we had to deal with ten acceleration in all

three directions (x, y, z), angular rates around all three axes (roll, pitch, and yaw) and

pressure sensors to obtain both altitude and airspeed.









The MAV128 R5 Power System

The initial goal of the power system for the MAV128 R5 was to separate the analog

and digital power systems in the hopes of improving the performance of the analog

processing system. Therefore, the MAV128R5A, shown in Figure 4-1, used separate

voltage regulators for the digital and analog systems. The TI REG113NA regulator used

in the MAV128 R4C for GPS power was utilized once again, since it was available in 5V

versions as well. One regulator powered all of the 5V digital systems and the altimeter

while both the gyroscopes and accelerometers each had their own 5V REG113 to provide

power. This was a precaution to try and keep the sensors isolated from one another. A

REG113 3.3V regulator once again provided power for the GPS. However, all of the

regulators received their input voltage direct from the battery. Though the REG113 was a

small device and not capable of dissipating more then a watt of thermal requirements, it

could still handle the 7.4V of the two-cell battery.

The exception was the AC4490. Due to the fact that it required almost 0.5A there

was no way a REG113 could provide it power. Furthermore, the LM3940 used

previously could not handle 7.4V as an input, as it was targeted only for converting 5V to

3.3V. A new voltage regulator had to be

found. The solution was the LMS8117A.

Also available in a SOT-223 package,

the device was barely capable of

dissipating the heat we needed.

However, we increased it's ability to

transfer heat to the rest of the board and
Figure 4-1: MAV128 R5A









the surrounding air by increasing the

size of the pad that it's main ground tab

is connected to, and then removing the

protective covering over that pad from

the PCB design (Figure 4-2).

We then sent out the MAV128

Figure 4-2: MAV128 R5A AC4490 Regulator R5A design to be fabricated. The first

boards came back and were assembled

right before we were told that we had to cut the weight of the flight system even more in

R5. However, this was extremely difficult. A few grams might be trimmed from the

MAV128, or we might possibly be able to find a smaller GPS unit (though using one

might result in decreased performance), but neither attempt would really make a

difference in the overall weight of the flight electronics. The 26-gram antenna used for

data transmission was a generic device obtained from Aerocomm, and had a rubber

shroud covering that was not necessarily needed for our application. We might have been

able to reduce the weight by getting a custom antenna designed, but even then we would

only be saving about 10 grams.

Therefore, our solution was to remove the dedicated electronics battery from the

equation, which took up almost half of the weight of the flight system. However, as a

result the electronics had to be powered off of the three-cell motor battery. The power

system had never been designed to be capable of handling a battery input of 12V. As a

test, we hooked up a 12V battery to the MAV128 R5A board, which had already been

verified to work at 7.4V. Within a minute all of the voltage regulators connected to the









battery went into thermal shutdown. We therefore had to completely redesign the power

system on the MAV128 R5A to support operation off of the motor battery.

We decided to use the same regulator for both the accelerometers and the

gyroscopes, as it didn't appear that any real performance gain was being produced from

separating the two sensors. Therefore, we needed two 5V regulators, one for analog

devices, and one for digital. The 3.3V regulator that was to provide power to the GPS

could exist underneath the digital 5V regulator, essentially acting as a second stage

device. We still had the issue of the 3.3V regulator for the AC4490. If it were a challenge

to enable a 5V regulator to receive 12V and source a large amount of current, then it

would be even more difficult to do so with a 3.3V regulator. Therefore, it was decided

that the voltage regulator for the AC4490 should not be connected to the battery directly,

but instead receive its power from another voltage regulator, essentially acting as a

second stage device.

Even with this layout, we still required several voltage regulators capable of

dissipating a lot of thermal energy, including both the analog 5V regulator since it would

have the battery for an input, and the AC4490 regulator since it would have to source a

large amount of current. The digital 5V regulator would have to handle both of these

situations, since it would be providing the AC4490 voltage regulator with power.

However, the only way that we could ensure that these devices could operate and not go

into thermal shutdown was to move to larger packages then the SOT-223 currently being

used. Using three such voltage regulators would take up a lot of real estate on the

MAV128 board.















4 4



,7 2

0
a. TO-263 PACKAGE
x PCB MOUNT
I SO. IN. COPPER


-40 -25 25 75 125
AMBIENT TEMPERATURE (OC)


MAXIMUM POWER DISSIPATION vs TEMPERATURE
3.0
Condition 1
25 ----- Condition2 -
----- Condition 3



"- "--..
205 """"-
20 __



-50 -25 0 25 50 75 100 125
Ambient Temperature (C)
CONDITION PACKAGE PCB AREA THETA J-A
1 MSOP-8 1 sq. n. Cu, 1 Side 71
2 MSOP-8 0.25 sqin Cu, Sie 90
3 SOT-23-8 None 200


Figure 4-3: TO-263 Maximum Power Figure 4-4: SOT-23 Maximum Power
Dissipation2 Dissiaption3



The solution was to use one single first stage voltage regulator to supply all current


for the other devices on the MAV128 R5B. It took in the 12V source from the battery and


produced 5.5V. The REG113 regulator, which has a dropout voltage of less then 400


mA, can easily produce 5V or 3.3V under these conditions. By moving to the


REG113EA, which used a MSOP package, there was an increase in size of 167%, but a


significant increase in thermal dissipation, as can be seen in Figure 4-4. Connecting all of


the ground pins of the device to a small ground plane that was then exposed to the air was


further insurance against the devices overheating. Under this setup, the REG113 could


power the GPS, digital, and analog systems.


The AC4490 Voltage Regulator, however, required far too much current to use the


REG113 Voltage Regulator, or even the LM3940 or LMS8117A used in previous


systems. Now that the input to the regulator was 5.5V, even the latter device was


incapable of dissipating the thermal energy resulting from providing power to the RF


2 LM1085 Datasheet, National Semiconductor, 2003


3 REG113 Datasheet, Texas Instruments, 2003


]I









Transceiver. Therefore, there were two

voltage regulators that needed to be

upgraded to a larger package. Using

the analysis in Table 4-1, we

determined that the solution was to use

devices from National's LM1085 and

LM1086 series of regulators. These

devices, which met voltage and current
Figure 4-5: MAV128 R5B
requirements, were available in the

TO-263 surface mount package. The package is over three times the size of the SOT-223

used previously, but it is also able to dissipate twice as much power as well (Figure 4-3,

Figure 3-5). Therefore, a LM1086 3.3V regulator, which could source up to 1.5A, was

used to power the AC4490. The LM1085 could source up to 3A, and was therefore well

suited to power the rest of the MAV128. Since the first stage regulator required a non-

standard output of 5.5V, an adjustable version of the LM1085 was used.

The evolution of the power system over the development of the Revision 5

MAV128 can be seen in Figure 4-7. With these changes, the Revision 5B was successful

at powering the system when connected to a three cell battery (Figure 4-5). The first stage

and RF transceiver regulators could dissipate the heat generated, though the temperature

of the devices increased to near 1000C. The devices themselves could handle this heat,

but they increased the temperature of the air around them significantly. The heat was too

much for the AC4490, which was connected to the MAV128 on the bottom of the board









and therefore right next to the

LM1086IS-3.3. Long-term

operation resulted in damage to the

RF transceiver, which no longer

communicated and had to be

replaced.

Because of this issue, the

layout of the power system was

changed for the MAV128 R5C. Figure 4-6: MAV128 R5C Power System: First
Stage and AC4490 Regulators
The connector for the AC4490 was

moved to the opposite side of the board, to keep the RF transceiver away from the heat of

the voltage regulators. Significant surface area of the MAV128 was also used to create

exposed copper planes connected to the main tab of the TO-263 regulators. (Figure 4-6)

This increased the ability of the devices to transfer heat to the board and ambient air, and

therefore lowered their internal temperature. This layout of the power system was

verified on the benchtop to successfully operate for long periods of time with out shutting

down due to heat. Though there is a detectable increase in temperature around the board,

it has never risen to the point where the AC4490 is affected. The MAV128 R5C (Figure

4-8) board has since been successfully flown on a 24" MAV, and the power system was

able to operate under flight conditions.













Battery Digital (5V)


MAV128 R4C


ll 2V Y LM10851S-ADJ
Analog (5V) GPS (3 3V)
REG113EA-5 REG113EA-33
MAV128 R5B, R5C


Figure 4-7: MAV128 Power Systems


Table 4-1: Analysis of Thermal Dissipation Issues in Powering the MAV128R5, GH-80,
and AC4490
Voltage Regulator Thermal Dissipation PT = I (VOUT VIN)
AC4490-500 (3.3V) Current Consumption 492 mA
MAV128R5 (5V) Current Consumption 10 mA
(Digital)
MAV128R5 (5V) Current Consumption 50 mA
(Analog)
GH-80 (3.3V) Current Consumption 88 mA
Three-Cell Battery, MAV128R5B/C, AC4490-500, GH-80
First Stage LM1085 5.5V (12V 5.5V) (492 mA + 10 mA + 50 mA + 88 mA
= 4.16W
R5 Digital REG113EA 5.0V (5.5V 5.0V) 10 mA = .005W
R5 Analog REG113EA 5.0V (5.5V 5.0V) 10 mA = .025W
AC4490 LM3940 3.3V (5.5V 3.3V) 492 mA = 1.0824W
GH-80 REG113EA- 3.3V (5.5V 3.3V) 88 mA= .1936W









Development of the Inertial and Analog Conversion Systems

While we were determining the best way to ensure that the MAV128 R5 could

survive on a three-cell battery, we were also focusing on developing the inertial

measurement system that needed to be integrated into the flight computer. The major

components that were required were accelerometers to measure the instantaneous

acceleration of the aircraft in all three directions, and gyroscopes to determine the angular

rates of movement about all three axes. The MPX4115A also needed to be migrated to

the new system. A request was also given to include in R5 a sensor to measure the

airspeed of the MAV. After some research, it was determined that the Motorola

MPXV4006 pressure sensor was being used for this purpose in other autopilots [23]. We

still needed to determine a way to accurately convert the analog signals of all of these

devices into data accurate enough to be used in our controller.

Since as always size was an issue, we decided to focus on new accelerometers and

gyroscopes developed using MEMS technology. Many IMUs that were intended for

small platforms used accelerometers from Analog Devices, including the 3DM-G. We

initially focused on using the ADXL210, which had a range of +10Gs. The ADXL210,

like most of the Analog Devices accelerometers, was dual-axis, and therefore we only

needed two sensors to measure acceleration in all three directions. The ADXL210

provided digital pulses for outputs, with the duty cycle determining the acceleration it

was measuring. However, the Megal28 only had two input capture signals, and so

therefore we were forced to use a different method for recording the data. Fortunately, the

ADXL210 could also provide an analog signal with the addition of a capacitor and an op-

amp buffer.









There were a few options for the gyroscopes. The 3DM-G used the ENC-03J

gyroscopes from Murata. However, we were informed that they were not available for

military applications. Tokin also had SMD gyroscopes available, but they were not

MEMS based. In the end, we again went with Analog Devices, using their ADXRS300

gyroscope. The device measures angular rates at up to 3000/s about one axis. One major

issue was the fact that the device was only available in a small ball grid array (BGA). It

was very difficult to attach this device to the board using the equipment we had available.

We had some success in using a SMD Soldering Station, focusing its hot hair on the

underside of the BGA as we lowered it onto the board, and allowing the solder on the

bottom of the device to flow onto the pads. However, this process was never completely

reliable.

Using these devices meant we needed two accelerometers and three gyroscopes to

cover all three axes. One accelerometer was placed directly on the main board, and

provide acceleration in the x and y directions. A gyroscope also was placed on the board

to provide the yaw rate. There were also two daughter boards, with one containing an

accelerometer and gyroscope to prove acceleration along the z-axis as well as the yaw

rate. The other daughter board contained just a gyroscope to provide the pitch rate.

We had designed the MAV128 R5A with the goal of keeping the accelerometer and

gyroscope sensors isolated from the digital power system. At the time we were also still

working on the MAV128 R4C, attempting to develop a circuit board using op-amps to

increase the sensitivity of the sensor. Therefore, like the Revision 4C, the new system

included a connector to a separate circuit for amplifying the altitude data.









By the time we started

developing the MAV128 R5B to

correct the issues in the power

system, we had also realized

through our experimentation with

the altitude sensor that part of the

problem was the affect the

Mega128 processor core had on

the internal ADC. Therefore, when Figure 4-8 MAV128 R5C

we redesigned the power system for the MAV128 R5B, we also made an attempt to

further isolate the analog systems from the digital. The ATmegal28 did provide separate

power and ground connections for the internal ADC. These signals were tied to analog

power and ground, as were all of the inertial sensors. As discussed earlier, the digital and

analog power systems used different voltage regulators. And starting with the R5B

system, we kept the analog and digital grounds separate from one another. As soon as

ground entered into the board at the main power connector, it connected directly to the

digital ground. It also connected through a OQ Resistor to Analog Ground. This was the

only connection between the two signals. This way, fluctuations on the digital ground

plane due to the processor operation were less likely to affect the analog systems.

Another change made during the development of the MAV128 R5B was a

modification of the layout of the inertial sensors. In the previous designs, one daughter

board included an accelerometer and gyroscope, and other board just had a gyro. To

reduce the complexity of the system and make manufacturing easier, a single daughter









board was designed containing both an accelerometer and a gyroscope. Two such

daughter boards were used with the system, and provide all of the inertial data except the

yaw rate. This was again covered by one gyroscope that remained on the main board.

We also went from using the ADXL210 accelerometers to the more accurate

ADXL203s, a new sensor from Analog Devices. It also provided analog outputs instead

of PWM signals, meaning the op-amp buffers were no longer needed. However, while the

ADXL210 had a range of 10Gs, the 203 only had a range of 1.7Gs. Some of the issues

with the 3DM-G in the earlier tests had been the saturation of its accelerometers under

certain flight conditions, and those devices had a similar range to the ADXL203.

However, the sensor was now being used on the PocketMAV Controller with some

success. Therefore, we decided to use these sensors for the MAV128 R5B as well. One

benefit of the new layout of the inertial sensors was that if there were problems with the

ADXL203, it would be easy to go back to the old accelerometers by just replacing the

daughter boards.

With the analog and digital systems completely isolated, we were hopeful that the

amplifier system developed previously would be enough to obtain a sensitivity of a few

feet from the MPX4115A. Therefore, the amplification circuit was migrated directly into

the MAV128 design. However, tests of the MAV128 R5B showed no further

improvement in the resolution of the altitude sensor. We were getting data from the

inertial sensors, but there was no real way to tell if the results were accurate enough for

autonomous control. It was therefore possible that the same problems we were facing

with the altitude sensor were occurring with the other sensors as well.









Because of these issues, we decided to search for a high resolution external ADC

that would interface to the Megal28. We hoped to further isolate the analog systems from

the Megal28, and achieve greater sensitivity in all of the sensor data. The MAV128 R5C

was designed to use this new ADC to convert the analog signals of the inertial sensors, as

well as the altitude and airspeed pressure sensors. The Megal28 ADC was still used for

low-resolution devices. This included the servo feedback systems, as well as a simple

resistor divider circuit to drop the battery to a level that could be read by the Megal28.

This allowed us to monitor the battery voltage and land the MAV if the battery was low.

Finally, the Megal28 also read a temperature sensor that was included in the gyroscope.

Since the inertial and pressure sensors were affected by temperature conditions, it was

decided to include the signal in case it was needed for future controller development.

Our requirements for an external ADC included a serial interface through either the

SPI Port or the I2C Bus. We did not want to deal with having to connect several traces

between the Megal28 and the ADC, which could make the layout of the board more

difficult, and so ADCs with parallel interfaces were not considered. We needed at least 8

channels to support the accelerometers, gyroscopes, and the pressure sensors, though the

sensors could potentially be divided into multiple ADCs if necessary. It was felt that

using a 14-bit or less ADC would most likely result in not enough sensitivity for our

needs based on the previous experiments. Furthermore, it was found that often companies

offer the same device in 12, 14, or 16 bit resolutions, with the only difference being the

cost of the devices. Therefore, we only looked for ADCs with at least 16 bits of

resolution.









In the end, we went with even higher resolution in choosing the TI ADS 1256, an 8

Channel 24-bit Serial ADC, for the MAV128 R5C. With this much accuracy, any issues

in the data were most likely to result from the sensors themselves or interference from

other devices, not the ADC. Hopefully, the MAV128 R5C design had already dealt with

these issues though the selection of the sensors and the isolation of the power systems.

The ADS 1256 interfaced to the Megal28 through the SPI Port. It included multiple

features, including a low pass filter, a programmable gain amplifier, and also an input

buffer that helped produce the large resolution for the ADC. However, the input buffer

could only function with inputs of less then 3V. The accelerometers and gyroscopes

produce 2.5V under null measurements, but could easily go above this limit.

Furthermore, the measurement of the altitude sensor on the ground was about 4. 1V. We

therefore had to disable the input buffer, though we only sacrificed one or two bits of

resolution as a result.

Once the details of interfacing the ADS 1256 to the inertial sensors and the

Megal28 had been determined, the MAV128 R5C design was sent to be fabricated.

When the boards were returned, the parts were soldered onto the system. However, our

method of attaching the gyro BGAs to the PCBs failed this time. The Yaw Gyroscope

only produced a signal when it was heated up by the air gun of the SMD soldering

station, and then only for a few minutes. Obviously, there was an intermittent connection

on the bottom of the device. The solution was to instead have two complete MAV128

R5C systems assembled by a contractor. This ensured that all of the components were

accurately placed on the boards.

























0P
21

`Pe
*s m & i
^ 4 Qa0
0 I-









With the MAV128 R5C
Raw and Filtered (10 Avg)
PCBs fabricated, assembled, and 15000
7500
returned to us, we set out to 5000
2500
modify the code so that not only 0'2 40 60 800 100

GPS data but also inertial Figure 4-10: MAV128 R5C Altitude Data

telemetry could be recorded and sent down to the ground station upon request. The major

change was the addition of the ads1256_int module, which used the Megal28's SPI Port

to control the ADS1256, access the inertial sensor data, and store it to the mav data

structure. The overall program was changed from dataMAV to autoMAV, because for the

first time since the early experiments of the prototype flight systems, we were going to

attempt to have an onboard controller fly the airplane directly through the MAV128. The

ground station was only to be used to monitor the situation.

Once we had finished developing the autoMAV program (Figure 4-9), we set out to

test the MAV128 R5C on the ground. We already knew that the sensors were sending

data, with the gyros showing no angular rates under static conditions. Furthermore, the

accelerometers could only detect the force of gravity. However, until flight-testing began

we did not know whether the ADS 1256 and isolated analog power system would be

enough to provide accurate inertial data for autonomous control. We did have the ability

to test the sensitivity of the altitude sensor though. We carried the MAV128 R5C board

up and down several flights of stairs, and then analyzed the telemetry using a low pass

filter to determine the sensitivity of the system. Both the raw and filtered data are shown

in Figure 4-10. We determined that the sensor now had a sensitivity of around 2.5 feet,

due to the fact that we could actually determine the discrete changes in altitude that









occurred for every step taken. It appeared that our issues in converting the analog data

were solved. We now had to test the system under actual flight conditions to determine

whether the MAV128 R5C could be used as an IMU to autonomously fly a MAV

Flight Testing and Onboard Controller Development

By the time we were ready to begin flight-testing of the MAV128 R5C, the second

version of the AVCAAF plane was ready (Figure 4-11). Therefore, the plane was

modified so that the servos could be controlled directly by the MAV128. This method of

control had not been attempted since the initial tests of the prototype flight system, and

had been abandoned because of controllability issues. However, this functionality was

necessary for developing an onboard controller, so our first flight tests were to verify the

operation of the "fly-by-wire" system.

We soon ran into issues, however, with the servos intermittently rotating the

control surfaces to maximum deflection. After some analysis, it was determined that the

servo pulse, which should occur at a frequency of 50 Hz with a duty cycle of 5% to 10%,

was occasionally not switching

correctly. The reason for this

was that the servo channels

were using the output compare

functions of the Mega128.

When a Timer in the Mega128

reached a point equal to the

compare register, the servo

Figure 4-11: AVCAAF 2.0 control pin toggled, and the









processor would interrupt so that an interrupt service routine (ISR) function could run.

This function would then update the compare register so that the system would activate

again on the next transition of the servo command signal. However, with so many

different interrupt-based systems now being used in the Megal28, occasionally an ISR

would not have a chance to run before the next transition, and so the servo pulse would

remain at its current level, causing the uncommanded movements.

The solution was to use the Megal28's PWM Channels, a different part of the

Timer System, so that the processor could generate the servo pulses without relying on

code. This had previously not been used because using the PWM system resulted in the

loss of an input capture device. However, the peripheral in question was not being used,

and it was necessary to move to PWM control if the fly-by-wire system was to work

properly. Therefore, the necessary changes to the servo_control module were made. The

move to PWM control eliminated the issues with the system, and the fly-by-wire system

has since been verified.

With the MAV128 now flying the airplane, we began sending inertial data to the

ground. With the addition of a 900 MHz gain antenna to the ground station, we were able

to improve the reception of the datalink, though still only at a rate of about 25 Hz. This

allowed us to have some view into the operation of the controller, which was now being

coded inside of the autoMAV control loop. The autoMAV code was modified to convert

the results from the ADS 1256 to floating point numbers representing the actual voltage

present at the inputs of the ADC. It could then be converted to the actual units of the

sensor in question, as shown in Table 4-2.









Due to the fact that each gyroscope varied in its null voltage by +.2V, autoMAV on

startup also averaged the first 10 initial readings of each gyroscope, average them

together, and used the results to calculate the null voltage of each device. However, this

also meant that on power up, the MAV128 R5C needed to remain stationary.

Table 4-2: Conversion Formulas for Inertial Sensors
ADS1256 ADC
OV 8388607 (0x7FFFFF)
2.5V 0
5V -8388608 (0x800000)
Float Voltage Representation Conversion 2.5V (A
2- (ADCrelt) + 2.5V
223 2--1
Inertial Sensors
Accelerometers
Null (OG Acceleration) 2.5V
Sensitivity 1 V/G
Gyroscopes
Null (0 / s Rotation) 2.5V
Sensitivity .005V/ o / s
Altitude Sensor
Null (0 Feet) z 4.1V
(ADC,. + 0.095
Conversion (Voltage to kPa) 5V
.009
Airspeed Sensor
ADCresult 0.045
Conversion (Voltage to kPa) P= 5V
.1533


A simple controller was developed by Mujahid Abdulrahim, and then ported into

the autoMAV program. A more advanced butterworth lowpass filter was used to further

filter the data. Then, the results from the accelerometers were used to determine the

gravitational vector, and from that data a state estimator could obtain the current pitch

and roll of the MAV. The controller then used a simple proportional controller to order

the servos to positions that would adjust the pitch and roll angles to 0, using the angular









rates from the gyros to control the movement. One initial result that came out of this

work was that we now knew for certain that the Megal28 could run an inertial-based

controller onboard. Whereas at minimum any control algorithm needs to run at about 30

Hz, the Megal28 was running the control loop at over 250 Hz.

Static tests in the lab showed that the state estimator could accurately determine the

orientation of the airplane. However, the first flight tests were unsuccessful in that the

MAV128 controller could not stabilize the MAV. Telemetry on the ground indicated that

the issue was the accelerometers, which were showing extremely noisy data, even with

the Butterworth low pass filter.

The accelerometers had a hardware low pass filter built in, with the cutoff

frequency determined by the value of an external capacitor at the output of the sensor.

We had initially used a .001 F capacitor, resulting in a 5 kHz bandwidth. As we were

seeing a lot of high frequency noise in the accelerometer signals, the capacitor was

changed to a ltF capacitor instead, giving us a cutoff frequency of 50 Hz. The noise still

remained, however. There was some though that the issues with the accelerometers was

that they were being affected by EMF generated by the MAV drive motor. However, it

was considered more likely to be caused by the vibrations of the airplane as it moved

through the air. The MAV was therefore hooked back up into the jig previously used in

the 3DM-G tests. Once again, these tests showed that vibration was a key factor, since the

accelerometer results rapidly oscillated whenever the plane was being shaken, whether

due to the drive motor or induced motion (Figure 4-12).










1.5


1


0.5l I I



1 84 16 49 1 914 9 7 1080 116 1246 1329 141; 1495 157 -AccelX
... .. AccelY
-1 AccelZ




-1.5' [l + 2 --------------------------------- --
-0.5 *'. -



-1.5


-2


Figure 4-12: Accelerometer Data


Adjustments to the cutoff frequency of the Butterworth filter had little affect. We

then came to the realization that the filter itself was not running properly. This was due to

the fact that the results of the Megal28-based filter differed from that of a PC-based filter

when the same raw data was supplied to both. We eventually determined that the issue

was with the frequency of the controller on the Megal28. The control loop ran as quickly

as possible, with the rate slightly varying as we inserted different instructions to try and

debug the system. Furthermore, the rate was also affected during those cycles where a

packet had been requested and was being sent to the ground station. We decide to modify

the program so that the control loop always ran at a rate of 50 Hz. The timer in the

Megal28 that was being used for capturing servo commands from the RC system was






68


used as a reference to stall the control loop at the end of a cycle until a full 20 ms had

passed. Once this modification was made, the results of the Megal28-based Butterworth

filter matched that of the ground. However, we still could not filter out the high

frequency oscillations in the data caused by the vibration of the MAV. AS a result, the

controller would not see the lower frequency changes in the gravitational vector resulting

from the movement of the plane. The filters we are employing are still unable to provide

us with the necessary data to allow the system to track the gravitational vector.
















CHAPTER 5
CONCLUSIONS AND FUTURE WORK

At this point we are still attempting to develop a simple onboard controller for the

MAV128 R5 that will enable inertial-based autonomous control. A few test flights have

been flown on a much larger RC aircraft. The resulting accelerometer telemetry, shown

in Figure 5-1, shows none of the high frequency noise we have noticed on the MAV

flights. Therefore, the issue with vibration appears to be directly related to the size of

platform that we must control. After investigation, we have discovered that we are not

alone in our difficulties to isolate the vibration of the aircraft. Other projects attempting


--AccelX
--AccelY
AccelZ


Figure 5-1: Accelerometer Results, 6 Foot RC Aircraft Platform









to develop autopilots for aircraft have often used a mixture of special mechanical mounts

to dampen the vibrations of the avionics system, as well as Kalman filters to retrieve the

necessary data from the accelerometers [24]. However, the former is dependent on the

design of the aircraft, and the latter must be dealt with from a controls perspective. At this

point, it appears that there is nothing more that can be done from the perspective of

hardware development to further develop the MAV128. Given the right mounting system

and the proper control methods, it will be possible for the MAV128 R5C to use its

integrated IMU to stabilize the MAV in the air, and use the GPS system already

developed in the Revision 4 platform to navigate.

With the combination of the MAV128 R5C Inertial and GPS Systems and the

vision processing capabilities of the ground station that are even now starting to come

online, the AVCAAF aircraft should be able to begin to perform increasingly complex

maneuvers in open fields. This would not, however, necessarily translate to being able to

do so in an urban environment. With the inertial based control onboard, the ground

station is no longer necessary for aircraft stabilization. Furthermore, if the control

algorithms for the GPS-based navigational systems are moved into the Megal28, the

ground station interface could be simplified to just provide status information about the

state of the aircraft, and allow the GPS waypoints used for navigation to be updated.

However, since the MAV128 will never be able to provide the computational power for

vision processing, this would have to remain in the ground station. As such, the MAV

would still need to send a video signal to the ground to be analyzed, and receive some

controller commands.









In the flight tests of the MAVs that have been conducted in an open field, video

noise occasionally occurs, but the system is often able to compensate. However, in an

urban environment, there may be several seconds or more of dropouts that could very

well be fatal for the survival of the airplane. It will be very difficult to ensure a reliable

link between the ground station and the MAV once the vehicle is flying amongst

buildings.

Therefore, the next step in the development of the avionics flight system must be to

start the process of migrating the vision processing functions of the ground station into

the MAV. The functionality of the MAV128 is of vital importance to the ability of a

MAV to survive in an urban environment, but it now must be relegated to one component

of an overall MAV flight system that will take data from all three systems inertial,

vision, and GPS and determine the necessary servo commands to continue its mission.

The complexity of this system will be far beyond that of the MAV128. The horizon

tracking system, the first component of the MAV vision processing system developed,

could successfully run on a 1 GHz processor only through the use of a hardware

framegrabber. A DSP, or a processor with special functions for image processing, would

most likely have to be used to have any chance of running the horizon tracking system

onboard. Even by reducing the complexity of the algorithm, we would still face immense

challenges in developing a traditional processing system that could accomplish this, given

the size, weight, and power limitations of a MAV platform. Furthermore, having horizon

tracking onboard will allow us to easily augment the inertial-based stabilization system

with vision and produce a more robust controller, Similarly, there are far more complex

vision processes that are just now coming online that must eventually be moved onboard









if we are to succeed in developing an autonomous MAV capable of urban operations [25,

26].

Fortunately, we have more options available to us then the traditional processing

systems. While initially the PLD market was focused on replacing simple logic systems

with a single device, the advances in Field Programmable Gate Arrays (FPGAs) over the

past few years has led to growing interest over the ability of the devices to function as

custom designs in multiple applications. The newest devices are so advanced that they are

now being used to not only augment DSP systems, but also replace them entirely.

Because of this, FPGA-based system-on-a-chip (SoC) technology is being using in many

applications, including vision processing. Just recently, FPGAs have been used in Sony

humanoid robots as a means for supporting stereo vision by running the necessary

algorithms to combine the two camera signals for processing [27]. The ability to develop

a custom digital system in an FPGA, as well as have this design perform multiple

operations in a single clock cycle, make FPGAs well suited for performing the intensive

pixel operations necessary in most vision processing applications.

Therefore, the next stage of flight computer development will be heavily focused

on designing an FPGA-based system capable of performing horizon tracking. It will also

have the processing power and capacity to handle more then just the vision-based

stability system, so that other components of the ground station vision systems may be

ported to the system. The same challenges faced in the development of the MAV128 in

the past year and a half will still be major factors in the development of this new system,

but emerging technology developed by the industry allowed us to move beyond those

challenges and develop a flight avionics system capable of autonomous control for such a






73


small aircraft. As such, new developments coming in the next year or so will also allow

us to add vision processing capabilities to the system, and move us closer to our goal of

developing an autonomous MAV capable of urban operations.















LIST OF REFERENCES


[1] J. M. McMichael and Col. M. S. Francis, "Micro Air Vehicles Toward a New
Dimension in Flight," World Wide Web,
http://www.darpa.mil/tto/mav/mav_auvsi.html, August 1997

[2] P. G. Ifju, S. Ettinger, D. A. Jenkins, Y. Lian, W. Shy, and M. R. Waszak,
"Flexible-wing based Micro Air Vehicles," 40th AIAA Aerospace Sciences Meeting,
Reno, NV, AIAA 2002-0705

[3] P. G. Ifju, S. Ettinger, D. A. Jenkins, and L. Martinez, "Composite Materials for
Micro Air Vehicles," SAMPE Journal, vol. 37, no. 4, pp. 7-13, July/August 2001

[4] S. Ettinger, M. C. Nechyba, P. G. Ifju, and M. Waszak, "Vision-guided Flight
Stability and Control for Micro Air Vehicles," Proc. IEEE Int. Conf. on
Intelligence Robots and Systems, vol. 3, pp. 2134-40, 2002

[5] S. Ettinger, M. C. Nechyba, P. G. Ifju and M. Waszak, "Vision-guided Flight
Stability and Control for Micro Air Vehicles," Advanced Robotics, vol. 17, no. 7,
pp. 617-40, 2003

[6] A. Kurdila, "Vision-Based Control of Agile, Autonomous Micro Air Vehicles and
Small UAVs in Urban Environments Project Overview," World Wide Web,
http://www.mil.ufl.edu/mav/presentations/kickoff mtg/overview.pdf, AVCAAF
Kickoff Meeting, UF-GERC, 27 October 2003

[7] L. Armesto, S. Chroust, M. Vincze, and J. Tornero, "Multi-rate Fusion with Vision
and Inertial Sensors," Proc. IEEE Int. Conf on Robotics and Automation, vol. 1,
pp. 193-99, 2004

[8] 3DM-G User Manual, Microstrain, Inc., World Wide Web,
http://www.microstrain.com/usermanuals/3DMGusermanual.pdf, April 2003

[9] Swift A2 GPS Receiver Product Specification, Axiom Navigation, Inc., Costa
Mesa, CA, 2002

[10] J. Grzywna, S. Kanowitz, P. Ifju, and M. Nechyba, "Integrating GPS with Micro
Air Vehicles," World Wide Web, http://www.mil.ufl.edu/-number9/mav/,
December 2002









[11] GPS Receiver Message Set Specification, Axiom Navigation Inc., Costa Mesa, CA,
2002

[12] CompactRF User's Guide, Microhard Corp., Calgary, AB, Canada, 2003

[13] MHX-910/2400 User's Guide, Microhard Corp., Calgary, AB, Canada, 2003

[14] ATmegal28(L) Datasheet, Atmel Corp., San Jose, CA, 2004

[15] J. W. Grzywna, D. MacArthur, J. Plew, and M. C. Nechyba, "Evaluation of a MAV
Inertial-based Flight Stability System using Vision Feedback," accepted to AIAA
Intelligent Systems Conference, Chicago, IL, September 2004.

[16] J. W. Grzywna, A. Jain, J. Plew, and M. C. Nechyba, "Rapid Development of
Vision-Based Control for MAVs through a Virtual Flight Testbed," submitted to
IEEE Int. Conf. on Robotics and Automation, Barcelona, Spain, April 2005.

[17] J. W. Grzywna, "A Flight Testbed With Virtual Environment Capabilities for
Developing Autonomous Micro Air Vehicles," Master's Thesis, University of
Florida, 2004

[18] Specifications for GPS Receiver Model GH-80, Furuno Electric Co., Ltd. System
Products Division, Camas, WA, 2002

[19] AC4490 Datasheet, Aerocomm, Lenexa, KS, 2003

[20] AC4490 User's Guide V1.7, Aerocomm, Lenexa, KS, 2004

[21] S. Jung, K. Lee, P. A. Barnswell, P. G. Ifju, J. W. Grzywna, J. Plew, A. Jain, and
M. C. Nechyba, "Vision-based Control for a Micro Air Vehicle : Part 1 : Testbed,"
submitted to AIAA Conf for Guidance, Navigation, and Control, Providence, RI,
August 2004.

[22] J. Kehoe, J. W. Grzywna, R. S. Causey, J. Plew, M. Abdulrahim, M. C. Nechyba,
and R. Lind, "Waypoint Navigation for a Micro Air Vehicle using Vision-Based
Attitude Estimation," submitted to European Micro Air Vehicle Conf,
Braunschweig, Germany, July 2004.

[23] Kestrel Autopilot 1.45 Description, Procerus Technologies, Provo, UT, 2004

[24] "autopilot: Do it yourself UAV," World Wide Web,
http://autopilot.sourceforge.net/, November 2004

[25] S. Todorovic and M. C. Nechyba, "A Vision System For Intelligent Mission
Profiles of Micro Air Vehicles," accepted by IEEE Trans. On Vehicular
Technology, In press






76


[26] S. Todorovic, M. C. Nechyba, and P. G. Ifju, "Sky/Ground Modeling for
Autonomous MAVs," Proc. IEEE Int. Conf Robotics and Automation, vol. 1, pp.
1422-7, September 2003.

[27] K. Sabe, M. Fukuchi, J.-S. Gutmann, T. Ohashi, K. Kawamoto, and T.
Yoshigahara, "Obstacle Avoidance and Path Planning for Humanoid Robots Using
Stereo Vision," Proc. IEEE Int. Conf. on Intelligence Robots and Systems, vol. 4,
pp. 3488-93, 2004















BIOGRAPHICAL SKETCH

Jason Plew was born in Bedford, IN, in 1979. He first began research into robotics

when entering high school in Palm Bay, FL. Upon graduating high school in 1998, Jason

began attending the University of Florida, where he started working at the Machine

Intelligence Lab. During this time, he has also worked through summer internships at

both P-Com, Inc., and Centerpoint Broadband Technologies, Inc. Jason later began to

work at Prioria Robotics, Inc., a small startup company located in Gainesville, FL. In

2003, he graduated from UF with Bachelor of Science degrees in both computer and

electrical engineering. He then began pursing a Master of Science in electrical

engineering, using the MAV research he was conducting as the basis for this thesis.




Full Text

PAGE 1

D E V E L O P M E N T O F A F L I G H T A V I O N I C S S Y S T E M F O R A N A U T O N O M O U S M I C R O A I R V E H I C L E B y J A S O N P L E W A T H E S I S P R E S E N T E D T O T H E G R A D U A T E S C H O O L O F T H E U N I V E R S I T Y O F F L O R I D A I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E U N I V E R S I T Y O F F L O R I D A 2004

PAGE 2

C opyr i ght 2004 by J a s on P l e w

PAGE 3

T hi s w or k i s de di c a t e d t o m y w i f e S ha l om ot h P l e w H e r l ove a nd s uppor t dur i ng t hi s r e s e a r c h a l l ow e d m e t o not onl y a c c om pl i s h m y go a l s f or t hi s p r oj e c t but a l s o e ns ur e d t ha t m y l i f e w a s f i l l e d w i t h ha ppi ne s s I t i s a l s o de di c a t e d t o m y f a t he r R i c ha r d P l e w w ho m a de a l l o f t hi s pos s i bl e I t w a s hi s s uppor t a n d gui da nc e t hr ough t he m a ny ye a r s o f hi gh s c hool r e s e a r c h pr oj e c t s t ha t e ve nt ua l l y l e d m e t o r obot i c s t he M a c hi ne I nt e l l i ge nc e L a b, a nd e ve nt ua l l y t he M A V p r oj e c t

PAGE 4

i v A C K N O W L E D G M E N T S I w oul d l i ke t o t ha nk t he p r of e s s or s a f f i l i a t e d w i t h t he M a c hi ne I nt e l l i ge nc e L a b, i nc l udi ng D r K e i t h D ot y D r A nt oni o A r r oyo, D r E r i c S c hw a r t z D r M i c ha e l N e c hyba D r K a r l G uge l a nd D r M i c ha e l L ync h, f or pr ovi d i ng m e w i t h t he know l e dge ne c e s s a r y f or t hi s r e s e a r c h. T he i r l e s s ons bot h i n t he c l a s s r oom a nd out s i de i t w e r e i nva l ua bl e I w oul d a l s o l i ke t o t ha nk t he m e m be r s of t he A V C A A F r e s e a r c h gr oup. T hi s o f c our s e i nc l ude s D r P e t e I f j u a nd hi s s t ude nt s f o r de ve l opi ng t he M i c r o A i r V e hi c l e s t ha t w e w oul d us e f or our p l a t f or m s W i t hout t he i r a m a z i n g ve hi c l e s none of t hi s w oul d ha ve be e n pos s i bl e T ha nks a l s o go out t o t he c ont r ol s t e a m l e d by D r A ndr e w K ur di l a a nd D r R i c k L i nd, w ho de ve l ope d t he na vi ga t i ona l c o nt r ol l e r f or t he M A V 1 28 R 4. A m ong t he i r s t ude nt s I w oul d e s pe c i a l l y l i ke t o t ha nk M uj a hi d A bdul r a hi m f o r de ve l opi ng t he i ne r t i a l ba s e d c ont r ol l e r f or t he M A V 128 R 5 t o ve r i f y i t s a bi l i t y t o c ont r ol a M A V a nd f or s ha r i ng hi s know l e dge a nd e xpe r i e nc e i n R C a i r c r a f t a nd c ont r ol s ys t e m s I w oul d a l s o l i ke t o t ha nk D r N e c hyba a nd t he o t he r s t ude nt s i n t he M I L M A V gr oup a nd m os t i m por t a nt l y J a s on G r z yw na f o r hi s de ve l opm e nt of t he gr ound s t a t i on s ys t e m a nd pa r t i c i pa t i on i n t hi s r e s e a r c h.

PAGE 5

v T A B L E O F C O N T E N T S pa ge A C K N O W L E D G M E N T S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i v L I S T O F T A B L E S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi i L I S T O F F I G U R E S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi i i A B S T R A C T . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x C H A P T E R 1 I N T R O D U C T I O N . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 R E V I S I O N 3: P R O T O T Y P E F L I G H T S Y S T E M . . . . . . . . . . . . . . . . . . . . . . . . . . 5 S ys t e m R e qui r e m e nt s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 F l i ght S ys t e m C om pone nt s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 I ne r t i a l S e ns or s : M i c r os t r a i n 3D M G . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 G l oba l P os i t i oni ng S ys t e m R e c e i ve r : A xi o m S w i f t A 2 . . . . . . . . . . . . . . . . . . . 8 R F T r a ns c e i ve r : M i c r oha r d M H X 2400 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 T he M A V 128 R 3 O nboa r d C om put e r . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 F l i ght S ys t e m a nd I nt e gr a t i on . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 P r ot ot ype F l i ght S ys t e m D e ve l opm e nt : C onc l us i ons . . . . . . . . . . . . . . . . . . . . . . . 24 3 R E V I S I O N 4: F L I G H T S Y S T E M W I T H O N B O A R D G P S . . . . . . . . . . . . . . . . . 26 F l i ght S ys t e m C om pone nt s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 G l oba l P os i t i oni ng S ys t e m R e c e i ve r : F ur uno G H 8 0D . . . . . . . . . . . . . . . . . . 27 R F T r a ns c e i ve r : A e r oc om m A C 4490 500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 T he M A V 128 R 4 a n d F l i ght S ys t e m I nt e gr a t i on . . . . . . . . . . . . . . . . . . . . . . . . . . 31 R e vi s i on 4 F l i ght S ys t e m C onc l us i ons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4 R E V I S I O N 5: F L I G H T S Y S T E M W I T H O N B O A R D I M U G P S A N D C O N T R O L L E R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 T he M A V 128 R 5 P ow e r S ys t e m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 D e ve l opm e nt of t he I ne r t i a l a nd A na l og C onve r s i o n S ys t e m s . . . . . . . . . . . . . . . . 55 F l i ght T e s t i ng a nd O nboa r d C ont r ol l e r D e ve l opm e nt . . . . . . . . . . . . . . . . . . . . . . . 63

PAGE 6

vi 5 C O N C L U S I O N S A N D F U T U R E W O R K . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 L I S T O F R E F E R E N C E S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 B I O G R A P H I C A L S K E T C H . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

PAGE 7

vi i L I S T O F T A B L E S T a bl e pa ge 3 1: A na l ys i s of T he r m a l D i s s i pa t i on I s s ue s i n P ow e r i ng t he R 4 F l i ght S ys t e m . . . . . . 37 3 2: P r e s s ur e S e ns or C onve r s i on D a t a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 3 3: F l i g h t S ys t e m s W e i ght D i s t r i but i on ( G r a m s ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4 1: A na l ys i s of T he r m a l D i s s i pa t i on I s s ue s i n P ow e r i ng t he M A V 128R 5, G H 80 a nd A C 4490 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4 2: C onve r s i on F o r m ul a s f or I ne r t i a l S e ns or s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

PAGE 8

vi i i L I S T O F F I G U R E S F i gur e pa ge 2 1: 3D M G I M U . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2 2: A xi om G P S R a di o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2 3: M H X 2400 R F T r a ns c e i ve r . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2 4: M A V 128 R 1, R 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2 5: M A V 128 R 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2 6: M A V 128 R 3 D a ught e r B oa r ds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2 7: R e vi s i on 3 F l i ght S ys t e m S of t w a r e A r c hi t e c t u r e . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2 8: 30" M A V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2 9: 3D M G B e nc ht op A na l ys i s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2 10: 3D M G B e nc ht op A na l ys i s w i t h D a t a S t a bi l i z a t i on . . . . . . . . . . . . . . . . . . . . . . . 23 2 11: M A V 128 F l i gh t C ont r ol l e r . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3 1: G H 80D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3 2: A C 4490 R F T r a ns c e i ve r . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3 3: R e v. 4 F l i ght S ys t e m S o f t w a r e A r c hi t e c t ur e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 3 4: M A V 128 R 4A R 4B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3 5: S O T 223 M a xi m um P ow e r D i s s i pa t i on . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 3 6: M A V 128 R 4C W i t h A C 4490 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 3 7: A V C A A F M A V P l a t f o r m 1 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3 8: R e vi s i on 4 F l i ght T e s t i ng S e t up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3 9: R e v. 4 F l i ght S ys t e m A ut onom o us W a ypoi nt N a vi ga t i on . . . . . . . . . . . . . . . . . . . 43

PAGE 9

i x 4 1: M A V 128 R 5A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 4 2: M A V 128 R 5A A C 4490 R e gul a t or . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 4 3: T O 263 M a xi m um P ow e r D i s s i pa t i on . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4 4: S O T 23 M a xi m um P ow e r D i s s i a pt i on . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4 5: M A V 128 R 5B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 4 6: M A V 128 R 5C P ow e r S ys t e m : F i r s t S t a ge a nd A C 4490 R e gul a t or s . . . . . . . . . . . . 53 4 7: M A V 128 P ow e r S ys t e m s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4 8: M A V 128 R 5C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 4 9: M A V 128 R 5C S of t w a r e A r c hi t e c t ur e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 4 10: M A V 128 R 5C A l t i t ude D a t a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 4 11: A V C A A F 2. 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 4 12: A c c e l e r om e t e r D a t a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 5 1: A c c e l e r om e t e r R e s ul t s 6 F oot R C A i r c r a f t P l a t f or m . . . . . . . . . . . . . . . . . . . . . . . 69

PAGE 10

x A bs t r a c t of T he s i s P r e s e nt e d t o t he G r a dua t e S c hool of t he U ni ve r s i t y of F l or i da i n P a r t i a l F u l f i l l m e nt o f t he R e qui r e m e nt s f or t he D e g r e e of M a s t e r of S c i e nc e D E V E L O P M E N T O F A F L I G H T A V I O N I C S S Y S T E M F O R A N A U T O N O M O U S M I C R O A I R V E H I C L E B y J a s on P l e w D e c e m be r 2004 C ha i r : A nt oni o A r r oyo M a j or D e pa r t m e nt : E l e c t r i c a l a nd C om put e r E ngi ne e r i ng T he goa l o f t hi s t he s i s w a s t o de ve l op a c om pl e t e a vi oni c s s ys t e m f or a M i c r o A i r V e hi c l e T hi s s ys t e m m us t s uppor t t he r e s e a r c h U F i s c ur r e nt l y c onduc t i ng i n de m ons t r a t i ng a n a ut onom ous M A V c a pa bl e of op e r a t i ons i n a n u r ba n e nvi r onm e nt T o s uppor t t hi s a bi l i t y a n onboa r d s ys t e m w a s ne e de d t ha t i nc l ude d bo t h a n I ne r t i a l M e a s ur e m e nt U ni t ( I M U ) f o r s t a bi l i z a t i on o f t he M A V a nd G P S f o r na vi ga t i ona l s uppor t T he r e w a s a l s o a ne e d f o r a n R F t r a ns c e i ve r f or l i n ki ng t he M A V t o a G r ound S t a t i on f or t e l e m e t r y a nd c ont r ol A f l i gh t c om put e r de s i gna t e d t he M A V 128, w a s r e qui r e d t o pr ovi de a n i nt e r f a c e be t w e e n a l l o f t he s e s ys t e m s a nd t o e ve nt ua l l y a l l ow f or a n onbo a r d c ont r ol l e r C om bi ne d w i t h a n onboa r d c a m e r a vi d e o t r a ns m i t t e r a nd vi s i on p r oc e s s i ng a l gor i t hm s on t he gr ound s t a t i on t hi s s ys t e m w a s t o a l l ow us t o be gi n de ve l opi ng a M A V c a pa bl e of t he a dva nc e d m a ne uve r s r e qui r e d f or f l yi ng a m ong bui l di ngs A pr ot ot ype f l i ght s ys t e m w a s de ve l ope d f o r a 24 M A V but c oul d not de m ons t r a t e a ut onom ous c ont r ol a nd na vi ga t i on due t o s i gni f i c a nt noi s e i n t he da t a f r om t he I M U a nd

PAGE 11

xi t he ove r a l l w e i ght of t he onboa r d s ys t e m I n t he de ve l opm e nt of t he ne xt s ys t e m t he w e i ght w a s r e duc e d a nd onboa r d a ut ono m ous na vi ga t i on w a s de m ons t r a t e d, w i t h f l i ght s t a bi l i z a t i on be i ng ha ndl e d by a gr ound ba s e d hor i z on t r a c ki ng s ys t e m a l r e a dy de ve l ope d. W e t he n i nt e g r a t e d a n I M U i n t o t he M A V 128, a nd t hough de ve l opm e nt i n bot h a m e c ha ni c a l m ount i ng s ys t e m f or t he s ys t e m a nd a dva nc e d c ont r ol f i l t e r s a r e r e qui r e d t o e ns ur e va l i d i ne r t i a l da t a f l i ght t e s t r e s ul t s ha ve be e n e nc our a gi ng. O nc e t hi s ve r s i on of t he f l i ght s ys t e m ha s de m ons t r a t e d a ut onom ous f l i ght w e w i l l t ur n t o i nc or por a t i ng i ne r t i a l vi s i on, a nd G P S c ont r ol onb oa r d, m ovi ng us e ve n c l os e r t o our goa l of de m ons t r a t i ng a ut onom ous ur ba n ope r a t i on of a M A V

PAGE 12

1 C H A P T E R 1 I N T R O D U C T I O N A m a j or f oc us of t he U ni t e d S t a t e s A i r F o r c e ( U S A F ) i n t he ne xt de c a de i s t he de ve l opm e nt of U nm a nne d A i r V e hi c l e s ( U A V s ) t ha t c a n be de pl oye d i n t a c t i c a l s c e na r i os a s oppos e d t o pur e s t r a t e gi c ope r a t i ons s uc h a s t he P r e da t or [ 1] T h i s i nc l ude s t he a bi l i t y f or s ol di e r s i n t he f i e l d t o de pl oy M i c r o A i r V e hi c l e s ( M A V s ) f or ons i t e s ur ve i l l a nc e T he m i l i t a r y a l s o w i s he s t o us e t he s e a i r c r a f t i n c om pl e x e nvi r on m e nt s s uc h a s ur ba n a r e a s i n m ul t i pl e s c e na r i os O f c our s e a s i s t he c a s e w i t h m os t m i l i t a r y t e c hnol ogy, t he r e a r e a l s o num e r ous a ppl i c a t i ons i n t he c i vi l i a n s e c t or T he M i c r o A e r i a l V e hi c l e L a b a t t he U ni ve r s i t y of F l or i da s D e pa r t m e nt of M e c ha ni c a l a nd A e r os pa c e E ngi ne e r i ng ha s be e n i nvol ve d i n t he r e s e a r c h a nd de s i gn of s uc h a i r c r a f t f o r s e ve r a l ye a r s a nd ha s be c om e ve r y pr of i c i e nt i n t he i r de ve l opm e nt o f M A V s [ 2, 3] T he i r t e c hnol ogy i n t he de ve l opm e nt of t he s e a i r p l a ne s ha s l e d t he m t o w i n t he I nt e r na t i ona l M A V c om pe t i t i on f o r t he pa s t f i v e ye a r s i n a r ow be c a us e of t he i r s m a l l s i z e a nd r e l a t i ve l y l ong r a nge T he a i r pl a ne s of t he M A V l a b ha ve t r a di t i ona l l y be e n c ont r ol l e d by o f f t he s he l f R C a i r pl a ne e qui pm e nt T o i nc r e a s e t he i r r a nge t he a i r pl a ne s of t e n f l e w w i t h a f or w a r d l ooki ng c a m e r a a nd a vi de o t r a ns m i t t e r a l l ow i ng t he pi l ot t o f l y t he a i r c r a f t w he n i t w a s be yond t he pi l ot s vi s ua l r a nge G i ve n t he s i z e of t he a i r p l a ne t he M A V pi l ot s ha d t o be c om e qui t e s ki l l e d t o ke e p t he pl a ne s i n t he a i r T he r e s e a r c h of S c ot t E t t i nge r he l pe d c ha nge t he g r e a t s ki l l r e qui r e m e nt of t he pi l ot s ; he de ve l ope d t he f i r s t a ut onom ous M A V s [ 4, 5] I n hi s r e s e a r c h he t ook a dva nt a ge of t he f a c t t ha t m a ny of t he M A V s w e r e s e ndi ng a vi de o s i gna l t o t he gr ound, a nd

PAGE 13

2 de ve l ope d a s ys t e m t o a na l yz e t he s e i m a ge s t o f i nd t he hor i z on A P I D c ont r ol l e r w a s de ve l ope d t o t a ke t he pos i t i on of t he hor i z on i n t he i m a ge a nd de t e r m i ne t he ne c e s s a r y c om m a nds t o t he s e r vos t o a dj us t t he c ont r ol s ur f a c e s t o ke e p t he hor i z on l e ve l I t c oul d a l s o ke e p t he hor i z on a t a s pe c i f i c a ngl e a l l ow i ng t he a i r c r a f t t o hol d a s pe c i f i c r ol l a ngl e a nd t hus or bi t a pos i t i on. T he ne c e s s a r y s e r vo c om m a nds w e r e t he n s e nt t hr ough a de vi c e t ha t w oul d c onve r t t he m t o s i gna l s unde r s t ood by a s t a nda r d R C c ont r ol l e r w hi c h w oul d t he n t r a ns m i t t he c om m a nds t o t he p l a ne ba c k t o t h e a i r pl a ne O nc e i t w a s w or ki ng i n r e a l t i m e t he vi s i on ba s e d f l i ght s t a bi l i t y s ys t e m c oul d ke e p a M A V i n t he a i r w i t hout a ny i nput f r o m a hum a n pi l ot F u r t he r w o r k l e d t o t he a bi l i t y f o r t he c ont r ol l e r t o t a ke i nput f r om a j oys t i c k, a l l ow i ng t he M A V t o be f l ow n by a ny unt r a i ne d pi l ot T he s uc c e s s of E t t i nge r s w or k l e d t o t he i ni t i a t i on of a ne w p r oj e c t A c t i ve V i s i on f or C ont r o l of A gi l e A ut ono m ous F l i ght ( A V C A A F ) [ 6] T he pur pos e of t hi s pr ogr a m w hi c h be ga n i n 2003 a nd i s t o l a s t f or f i v e ye a r s i s t o de ve l op M i c r o A i r V e hi c l e s t ha t a r e c a pa bl e of a ut onom ous f l i ght w i t hi n c om pl e x e nvi r onm e nt s s uc h a s ur ba n s e t t i ngs H ow e ve r t he r e a r e m ul t i pl e c ha l l e nge s t ha t m us t be ove r c om e t o a c c om pl i s h t hi s goa l T h ough a ut onom ous f l i ght h a d be e n de m ons t r a t e d t hr ough t he ho r i z on t r a c ki ng s ys t e m t hi s a l one w a s not s uf f i c i e nt f or pe r f or m i ng t he c om pl e x m a ne uve r s r e qui r e d i n a ny e nvi r onm e nt be yond a n ope n f i e l d s uc h a s a f or e s t or c i t y. T o be a bl e t o a c hi e ve t he c ont r ol ne c e s s a r y t o f l y a M A V a ut onom ous l y i n a n ur ba n e nvi r on m e nt w e ne e de d t o t a c kl e t he p r obl e m a t t h r e e l e ve l s T r a di t i ona l l y c ont r ol of a e r i a l r obot s ha s be e n a c c om pl i s he d t hr ough a n I ne r t i a l M e a s ur e m e nt U ni t ( I M U ) S uc h de vi c e s w hi c h of t e n i nc l ude a c c e l e r om e t e r s a nd gyr os c ope s a r e ve r y good a t de t e r m i ni ng t he i ns t a nt a ne ous m ove m e nt o f t he a i r pl a ne T he r e f or e t he y c a n be ve r y

PAGE 14

3 a c c ur a t e a t t r a c ki ng t he m ove m e nt of t he ve hi c l e o ve r a s hor t pe r i od o f t i m e H ow e ve r ove r l onge r pe r i ods of t i m e t he e r r o r of t he de vi c e s i nc r e a s e s t o t he poi nt t ha t t he y a r e i l l s ui t e d t o m e a s ur i ng t he m ove m e nt of t he a i r p l a ne G P S ha s of t e n be e n us e d t o pr ov i de t he ne c e s s a r y da t a t o r e c a l i br a t e t he I M U da t a A c ont r ol l e r c a n s t a r t w i t h a know n pos i t i on f r om t he G P S a nd t r a c k t he ve hi c l e s pr o gr e s s w i t h i ne r t i a l s e ns or s upda t i ng t he pos i t i on t o c or r e c t f or e r r o r s on e ve r y G P S pos i t i o n upda t e H ow e ve r t he G P S uni t c a n onl y de t e r m i ne t he pos i t i on of t he a i r c r a f t i n r e l a t i ons hi p t o t he e a r t h a nd pot e nt i a l l y pr e vi ous l y know n f i xe d obj e c t s i f s uc h da t a i s a va i l a bl e I t i s una bl e t o pr o vi de a ny r e a l i nf or m a t i on a bout t he c ha ngi ng e nvi r onm e nt a r oun d t he a i r pl a ne T he vi s i on s ys t e m how e ve r c a n be pl a c e d d i r e c t l y be t w e e n t he I ne r t i a l S ys t e m a nd t he G P S a s ye t a not he r s e ns or t o pr ovi de da t a t o t he a i r pl a ne c ont r ol s ys t e m A vi s i on s ys t e m s a bi l i t y, f or e xa m pl e t o t r a c k a m ovi ng t a r ge t or de t e r m i ne t ha t a n obs t a c l e i s a he a d of t he a i r c r a f t a l l ow s i t t o p r ovi de i n f or m a t i on t ha t w oul d not be a va i l a bl e t o t he m o r e t r a di t i ona l m e t hods of a i r pl a ne c ont r ol B y us i ng G P S V i s i on a nd a n I M U t he i na de qua c i e s of a ny one s ys t e m a r e c ove r e d by t he ot he r t w o [ 7 ] T he f i r s t c om pone nt of t he vi s i on s ys t e m ha d a l r e a dy be e n de ve l ope d, i n t he f or m of t he H o r i z on T r a c ki ng S ys t e m W hi l e t he 900 M H z x86 ba s e d c om put e r us e d i n t he or i gi na l t e s t s w e r e a de qua t e f o r r unn i ng t he ho r i z o n t r a c ki ng s ys t e m t he t r a ns f e r of t he hor i z on t r a c ki ng p r ogr a m t o ne w c om put e r t e c hnol ogy e ns ur e d t ha t e nough p r oc e s s i ng pow e r w a s a va i l a bl e f or m or e c om pl i c a t e d vi s i on p r oc e s s i ng t a s ks T he t e c hnol ogy f o r G P S a nd I M U s e ns or s t ha t c o ul d be us e d f o r a i r c r a f t on t he s c a l e of e ve n t he l a r ge s t of M A V s w a s onl y i n t he i ni t i a l l e ve l s of de ve l opm e n t N o of f t he s he l f s ys t e m s e xi s t e d t ha t c oul d be bot h f l ow n on M A V a nd i nt e gr a t e d w i t h t he vi s i on s ys t e m a l r e a dy be i ng

PAGE 15

4 de ve l ope d. T he r e f o r e w e ha d t o de ve l op t he t e c hn ol ogy our s e l ve s i e de ve l op a c om pl e t e f l i ght s ys t e m c a pa bl e of de m ons t r a t i ng a ut onom ous f l i ght s W e be ga n de ve l opi ng s uc h a f l i ght s ys t e m i n t he s pr i ng of 20 03. W hi l e t he D e pa r t m e nt o f D e f e ns e de f i ne s a M i c r o A i r V e hi c l e a s a n a i r pl a ne w i t h a w i ngs pa n of l e s s t he n s i x i nc he s t he c ur r e nt l e ve l of t e c hnol ogy w a s not a dva nc e d e nough t o be us e d i n s uc h a pl a t f o r m T he r e f or e w e c ont i nue d t o us e t he 24 i nc h w i ngs pa n s ys t e m s us e d i n t he e a r l i e r vi s i on ba s e d f l i ght s t a bi l i t y e xpe r i m e n t s T w o i ni t i a l a t t e m pt s t o de ve l op a f l i ght s ys t e m f o r a M A V w e r e f oc us e d on i de nt i f yi ng t he p r ope r de vi c e s t o us e a nd t o t he n de ve l op a m e a ns of t r a n s f e r r i ng t e l e m e t r y f r o m t he a i r pl a ne t o t he gr ound. T he s e s ys t e m s a l s o he l pe d us t o unde r s t a nd t he pr obl e m s w e f a c e d i n de ve l opi ng a M A V w i t h t he c a pa bi l i t y of pe r f o r m i ng a ut onom ous ur ba n ope r a t i ons w hi c h l e d us t o t he r e qui r e m e nt s f or t he t hi r d r e vi s i on of t he f l i gh t ha r dw a r e T hi s t hi r d s ys t e m c ul m i na t e d i n our f i r s t a t t e m p t a t a ut ono m o us f l i ght

PAGE 16

5 C H A P T E R 2 R E V I S I O N 3: P R O T O T Y P E F L I G H T S Y S T E M S ys t e m R e q u i r e m e n t s T he goa l s of t he f i r s t ye a r o f t he A V C A A F pr oj e c t w e r e ba s e d on t he r e a l i z a t i on t ha t t he a ut onom ous M A V s r e qui r e d I ne r t i a l a nd G P S c a pa bi l i t i e s i n a ddi t i on t o vi s i on. T he pr i m a r y e f f or t w a s t o de ve l op a n onboa r d f l i g ht s ys t e m f or a t w o f oo t w i ngs pa n M A V t ha t i nc l ude d bot h a n I M U a nd a G P S r e c e i ve r T hi s s e t up w a s de ve l ope d t o w or k i n c onj unc t i on w i t h a ne w g r ound s t a t i on i nc l udi ng a n e nha nc e d vi s i on gui da nc e s ys t e m ba s e d on E t t i nge r s w or k. T he ove r a l l s ys t e m r e qui r e m e nt w e r e t o a ut onom ous l y f l y t he M A V t hr ough m ul t i pl e G P S w a ypoi nt s a nd be r e a dy f or de m ons t r a t i on a t E gl i n A F B i n J ul y 2003. I ni t i a l l y, t he de s i r e w a s t o bui l d a s m uc h pos s i bl e f unc t i ona l i t y i nt o t he s ys t e m a s pos s i bl e T hi s i nc l ude d t he a bi l i t y t o bot h s e nd s e ns or da t a t o t he gr ound a s w e l l a s s t or e i t i n t he onboa r d s ys t e m f or l a t e r r e t r i e va l M ul t i pl e l oc a t i ons of t he c ont r ol l e r ( bot h on t he gr ound s t a t i on a nd i n t he onboa r d s ys t e m ) w e r e c ons i de r e d. W e a l s o w a nt e d t he a bi l i t y t o c ont r ol t he s e r vos t hr ough bo t h t he onbo a r d f l i ght s ys t e m a nd t hr ough s t a nda r d R C e qui pm e nt a l l ow i ng t he c om put e r t o c ont r ol t h e a i r pl a ne a nd ye t i ns ur e t ha t a hum a n pi l ot c oul d s t e p i n i f ne c e s s a r y. A s e xpe r i m e nt a t i o n be ga n, s om e of t hi s f unc t i ona l i t y w a s de t e r m i ne d t o be unne c e s s a r y. O t he r c om pone nt s p r ove d t o be i na de qua t e f or a c c om pl i s hi ng t he t a s k of de ve l opi ng a c ont r ol l e r f or t he a i r pl a ne a nd w e r e a ba ndone d. A s a r e s ul t t he pr ot ot ype f l i ght s ys t e m t ha t w a s de ve l ope d by t he e nd of t he s um m e r o f 2003 ha d gone t h r ough s e ve r a l i t e r a t i ons a nd t he r e l a t i ons hi p be t w e e n t he di f f e r e nt c om pone nt s of t e n c ha nge d.

PAGE 17

6 T he r e w e r e m u l t i pl e c ha l l e nge s i n de ve l opi ng t he onboa r d s ys t e m T he m a j or i s s ue w a s t he w e i ght c ons t r a i nt A M A V w i t h a w i ngs pa n of t w o f e e t o nc e ou t f i t t e d w i t h m ot or s s e r vos ba t t e r i e s a nd R C e qui pm e nt c oul d onl y ha ndl e pa yl oa ds of l e s s t he n 100 gr a m s F ur t he r m or e t he r e w a s t he di f f i c ul t y i n de v e l opi ng a s ys t e m t ha t w a s l oc a t e d onboa r d t he a i r p l a ne a c c e s s t o t he de vi c e w he n f l yi ng i n t he a i r w a s l i m i t e d, a nd i t w oul d be di f f i c ul t t o r e pl i c a t e t he s e c ondi t i ons i n t he l a b. T hi s w a s e s pe c i a l l y t he c a s e w he n i t c a m e t i m e t o de ve l op a c ont r ol l e r t o l i nk t he onboa r d s e ns or s t o t he c ont r ol s ur f a c e s of t he M A V I ni t i a l l y, i t w a s hope d t ha t a c ont r ol l e r c oul d be de ve l ope d t o f unc t i on onboa r d t he a i r c r a f t A nd a s t he r e s e a r c h pr ogr e s s e d, t hi s r e m a i ne d t he e nd goa l of t he f l i ght s ys t e m H ow e ve r a ny t i m e a n onboa r d c ont r ol l e r w a s i n n e e d of s e r i ous m odi f i c a t i on, e i t he r w e w oul d ha ve t o l a nd t he a i r p l a ne s o t ha t t he onboa r d s ys t e m c oul d be r e pr og r a m m e d, or w e w oul d ha ve t o de ve l op a r obus t m e t hod of t r a ns m i t t i ng a ne w c ont r ol l e r t o t he a i r pl a ne S i nc e bo t h of t he s e opt i ons w e r e c ons i de r e d t o be uns ui t a bl e f or t he de ve l opm e nt pha s e w e de c i de d t o de ve l o p t he c on t r ol l e r on a g r ound s t a t i on. T hi s a ppr oa c h m a de i t e a s i e r f or t he c ont r ol s t e a m t o de ve l op t he ne c e s s a r y a l gor i t hm s t o f l y t he a i r pl a ne a ut onom ous l y, a l l ow i ng t he f us i on of t he vi s i on, i ne r t i a l a nd G P S da t a t o oc c ur i n a s i ngl e l oc a t i on. T he a i r p l a ne ne e de d t o s e nd a l l of i t s s e ns or da t a t o t he gr ound s t a t i on, a nd e i t he r pr oc e s s s e r vo c om m a nds t hr oug h s t a nda r d R C e qui pm e nt or t hr ough t he onboa r d f l i ght s ys t e m O nc e t he pr oc e s s of de v e l opi ng a c ont r ol l e r f o r t he M A V on t he gr ound ha d be e n r e f i ne d w e be ga n w or k on t h e de ve l opm e nt of t r a ns f e r r i ng t he s ys t e m t o t he onboa r d ha r dw a r e

PAGE 18

7 F l i gh t S ys t e m C om p on e n t s T he m a j or goa l of t he f l i ght s ys t e m w a s t o ha ve on boa r d i ne r t i a l s e ns or s a nd a G P S r e c e i ve r D ue t o w e i ght r e s t r i c t i ons ve r y s m a l l s e ns or s ha d t o be f ound, e ve n i f t he t r a de of f w a s i n r e duc e d a c c ur a c y. F ur t he r m o r e w e ne e de d t o i nc l ude t he a bi l i t y f or t he f l i ght s ys t e m t o t r a ns m i t a t l e a s t s om e of t hi s s e ns or da t a t o t he gr o und. F i na l l y, s om e f or m o f a p r oc e s s i ng s ys t e m w a s r e qui r e d t o t i e a l l t h e s e i ndi vi dua l c om pone nt s t oge t he r a nd l i nk t he m t o t he s e r vo m ot or s c ont r ol l i ng t he p l a ne I n e r t i al S e n s or s : M i c r os t r ai n 3D M G I ns t e a d of a t t e m pt i ng t o bui l d a n I M U f r om di s c r e t e pa r t s w e t r i e d t o obt a i n a n i nt e gr a t e d c om pone nt l i ght i n w e i ght ye t a bl e t o gi ve us t he a c c ur a t e da t a ne c e s s a r y f or s t a bl e f l i ght T he s ol ut i on w a s t he 3D M G a n I M U f r om M i c r os t r a i n ( F i gur e 2 1) T h i s de vi c e i nc l ude d a c c e l e r om e t e r s i n a l l t h r e e di r e c t i o ns t o gi ve us t he i ns t a nt a ne ous m ove m e nt of t he a i r pl a ne a s w e l l a s gyr os c ope s f o r a l l t hr e e a xe s t o p r ovi de r o l l pi t c h a nd ya w r a t e s A l s o i nc l ude d w e r e m a gne t om e t e r s t o gi ve t he o r i e nt a t i on o f t he a i r c r a f t i n r e l a t i ons hi p t o t he E a r t h s m a gne t i c f i e l d a nd ot he r a s s or t e d s e ns or s A n onboa r d c hi p w oul d pr oc e s s t he s e ns or da t a t o pr ovi de not onl y f i l t e r e d s e ns or da t a but a l s o t he or i e nt a t i on r e s ul t s ( e g. E ul e r A ngl e s ) t ha t a r e us e f ul f or c ont r ol l e r s T he 3D M G c oul d be c om m uni c a t e d t o t hr ough a U ni ve r s a l A s ync hr onous R e c e i ve r T r a ns m i t t e r ( U A R T ) a s t a nda r d c om m uni c a t i ons de vi c e f ound on m os t F i gur e 2 1 : 3D M G I M U

PAGE 19

8 e m be dde d pr oc e s s or s a s w e l l a s on t he s e r i a l po r t s of pe r s ona l c om put e r s [ 8] G l ob al P os i t i on i n g S ys t e m R e c e i ve r : A xi o m S w i f t A 2 T he ot he r m a j or s ys t e m w a s t he G P S uni t O nc e a g a i n w e i ght w a s a m a j or i s s ue i n de c i di ng on w hi c h de vi c e t o s e l e c t H ow e ve r i n t h i s c a s e t he a c c ur a c y of t he da t a w a s not a s i m por t a nt a f a c t or N o G P S r e c e i ve r c ur r e nt l y a va i l a bl e c oul d r e t u r n da t a a c c ur a t e e nough t o be us e f ul i n t he f l i ght s t a bi l i t y p r obl e m a nd t he r e f or e t he G P S w a s onl y us e d f or na vi ga t i on B e c a us e of t hi s a G P S uni t t ha t w a s onl y a c c ur a t e t o w i t h i n s e ve r a l f e e t w a s not a n i s s ue T o s a ve on w e i ght w e l ooke d f o r a G P S w i t h a n i n t e gr a t e d a nt e nna T hough us i ng a n i nt e gr a t e d a nt e nna r e s ul t s i n l e s s c a pa bi l i t y i n d e t e c t i ng t he G P S s a t e l l i t e s e s pe c i a l l y i n e xt r e m e m a ne uve r s w e de c i de d t ha t t he s a vi ngs i n w e i ght m a de s uc h a c hoi c e w or t h t he de gr a da t i on i n na vi ga t i on a l da t a T he r e s ul t i ng s ol ut i on w a s t he A xi om S w i f t A 2 G P S R e c e i ve r [ 9] T hi s de vi c e s how n w i t h a n i nt e r f a c e i n F i gur e 2 2, c om m uni c a t e s bot h t he pos i t i on a nd c our s e of t he a i r c r a f t t h r ough a U A R T a nd ha d be e n us e d i n pr e vi ous w or k t ha t f oc us e d on e xpl or i ng t he pos s i bi l i t y of i n t e gr a t i ng t he hor i z on t r a c ki ng s ys t e m w i t h G P S [ 10 ] G i ve n t he r e c e i ve r s l ow w e i ght of 20 g r a m s a nd t he f a c t t ha t w e a l r e a dy ha d s om e of t he s e uni t s a va i l a bl e m a de t hi s c hoi c e f or a G P S uni t ne a r l y i de a l B y de f a ul t t he S w i f t A 2 c om m uni c a t e s us i ng t he i ndus t r y s t a nda r d N M E A P r ot oc ol w hi c h c a n p r ovi de m ul t i pl e na vi ga t i ona l da t a i nc l udi ng F i gur e 2 2 : A xi om G P S R a di o

PAGE 20

9 L ongi t ude L a t i t ude a nd C our s e S pe e d a nd B e a r i n g. T he dr a w ba c k w a s t ha t t he N M E A pr ot oc ol us e s A S C I I t o t r a ns f e r da t a w hi c h t a ke s u p m uc h m or e ba ndw i dt h t he n us i ng a r a w bi na r y f or m a t T he onboa r d s ys t e m f i r s t ha d t o c onve r t t he G P S da t a i nt o bi na r y i f i t w a s t o be s e nt t o t he gr ound W hi l e t hi s w a s f e a s i bl e w e w i s he d t o a voi d ha vi ng t o w a s t e pr oc e s s i ng t i m e pa r s i ng t he A S C I I s t r e a m F or t una t e l y, t he S w i f t A 2 c oul d a l s o be s e t t o c om m uni c a t e us i ng t he r a w S i R F s t a nda r d p r ot oc ol T he i nt e r f a c e t o t he G P S us i ng t hi s s t a nda r d pr ove d t o be m uc h s i m pl e r O ne m i nor i s s ue w a s t ha t t he S i R F pr ot oc ol unl i ke N M E A c om m uni c a t e d pos i t i on da t a us i ng a 3D a x i s s ys t e m w i t h t he or i gi n a t t he c e nt e r of t he e a r t h t he z a xi s poi nt i ng t o t he nor t h pol e a nd t he x a xi s a l ong t he e qua t or i a l pl a ne a nd pe r pe ndi c ul a r t o t he p r i m e m e r i di a n [ 11 ] F o r t he de ve l opm e nt of na vi ga t i ona l c ont r ol l e r t o be s t r a i ght f or w a r d t he f o r m a t of t he pos i t i on ne e de d t o be i n t he l a t i t ude l ongi t ude a nd a l t i t ude f o r m a t H ow e ve r t he c onve r s i on pr ove d t o be j us t a m a t t e r of c onve r t i ng f r o m t he C a r t e s i a n c oor di na t e s ys t e m t o a S phe r i c a l P ol a r s ys t e m a pr oc e s s s i m pl e t o c a r r y out on t he g r o und s t a t i on c om put e r R F T r an s c e i ve r : M i c r oh ar d M H X 2400 W e ne e de d s om e w a y of be i ng a bl e t o c om m un i c a t e w i t h t he f l i ght s ys t e m w he n i t w a s i n t he a i r T he ba ndw i dt h c oul d r a nge be t w e e n a f e w byt e s f or s e ndi ng s e r vo c om m a nds up t o t he p l a ne ( bypa s s i ng t he R C R a di o S ys t e m ) a nd ge t t i ng s i m pl e s t a t us i nf or m a t i on ba c k t o r e c e i vi ng s e ve r a l byt e s of t e l e m e t r y i nc l udi ng I ne r t i a l a nd G P S da t a T he ne c e s s a r y da t a ha d t o be s e nt a t a r a t e of 30 H z f or t he c ont r ol l e r on t he gr ound s t a t i on t o f unc t i on pr ope r l y W e a l s o ha d t o be c on c e r ne d w i t h t he r a nge of t he t r a ns c e i ve r W e ne e de d t o m a i nt a i n t he r a di o l i nk f or a t l e a s t a f e w m i l e s O ne be ne f i t of t he M A V s ys t e m w a s t ha t w i t h one of t he t r a ns c e i ve r s be i ng i n t he a i r w e w e r e ope r a t i ng i n c l os e t o L i ne of S i ght ( L O S ) c ondi t i ons

PAGE 21

10 W e i ght a nd s i z e w e r e a s a l w a ys a c onc e r n i n de t e r m i ni ng t he r e qui r e m e nt s f or a t r a ns c e i ve r W e a l s o ne e de d t o l ook f o r a de vi c e t h a t ope r a t e d i n one of t he unl i c e ns e d ba nds of s pe c t r um s o t ha t w e w oul d not ha ve t o s p e nd t i m e ge t t i ng a ppr ova l f r om t he ne c e s s a r y gove r nm e nt a ge nc i e s t o ope r a t e t he s ys t e m T he s e f r e que nc i e s i nc l ude d 900 M H z 2. 5 G H z a nd 5 G H z T he l a t t e r s pe c t r um w a s onl y j us t be gi nni ng t o be u t i l i z e d, a nd s o t he opt i ons f o r us i ng a de vi c e a t t ha t f r e que nc y ba nd w e r e l i m i t e d. W e t h e r e f or e de c i de d t o us e e i t he r a 900 M H z or 2. 5 G H z P r e l i m i na r y t e s t s s how e d t ha t us i ng a da t a t r a ns c e i ve r a nd a vi de o t r a ns m i t t e r ope r a t i ng a t t he s a m e f r e que nc y r e s ul t e d i n t oo m uc h vi de o noi s e on t he g r ound. W e di d ha ve t he a bi l i t y t o t r a ns m i t vi de o a t e i t he r 2. 5 G H z or 900 M H z s o i t w a s a n e a s y m a t t e r t o ke e p t he t w o de vi c e s f r om i nt e r f e r i ng w i t h one a not he r W e de c i de d t ha t i t w oul d t a ke t oo l ong t o de ve l op our ow n t r a ns c e i ve r e ve n us i ng t he R F c hi ps e t s s uc h a s I nt e r s i l s P r i s m t ha t ha ve b e c om e popul a r i n t he pa s t f e w ye a r s T he r e f or e a n o f f t he s he l f s ol ut i on w a s i nve s t i ga t e d. I ni t i a l w or k ha d be e n unde r t a ke n i n ge t t i ng da t a t o t he gr ound f r o m s om e of our i ni t i a l f l i ght s ys t e m s us i ng t he C om pa c t R F R a di o T r a ns c e i ve r f r om M i c r oha r d C or p T hi s de v i c e o pe r a t e s a t 900 M H z a nd c a n, unde r opt i m a l c ondi t i ons t r a ns f e r da t a a t up t o 19. 2K bi t s f or a r a nge of 20 m i l e s [ 12] T he or e t i c a l l y, w e s houl d t he n be a bl e t o s e nd a pa c ke t of 50 byt e s a t a r a t e of a bout 20 H z H ow e ve r e ve n w i t h t he r a t he r i de a l c ondi t i on s of ha vi ng one of t he r a di os i n t he a i r w e w e r e onl y a bl e t o s e nd i ne r t i a l da t a t o t he gr ou nd a t a r ound 10 H z S i nc e t hi s di d not m e e t our t a r ge t e d da t a r a t e of 30 H z a not he r r a di o ne e de d t o be f ound

PAGE 22

11 A f t e r a n e xha us t i ve s e a r c h f or ot he r of f t he s he l f s ol ut i ons w e s e t t l e d f or a not he r r a di o m a de by M i c r oha r d. T he M H X 2400 ( F i gur e 2 3) a dve r t i s e d a 115. 2K da t a t r a ns f e r r a t e a t t he s a m e r a nge a s t he C om pa c t R F r a di o. F or t ha t m a t t e r t he y of f e r e d t he M H X 900 r a di o a s w e l l w hi c h w a s i de nt i c a l t o t he M H X 240 0 e xc e pt t ha t i t ope r a t e d a t 900 M H z [ 13] U s i ng t he M i c r oha r d r a di os w a s a t t r a c t i ve be c a us e i t ga ve us f l e xi bi l i t y i n de t e r m i ni ng w hi c h f r e que nc y ba nds t he D a t a l i nk a nd V i de o s ys t e m s w oul d u s e H ow e ve r a m a j or dr a w ba c k w a s t he s i z e a nd w e i g ht w he r e a s t he C om pa c t R F w a s 2 x 1. 5 a t 20g t he M H X 2400 w a s 3. 5 x 2 a t 75g. H ow e ve r t he r e a ppe a r e d t o be no ot he r vi a bl e opt i ons a t t ha t t i m e a nd s o t hi s de vi c e w a s c hos e n t o a l l ow t e l e m e t r y t o b e s e nt f r om t he M A V t o a c ont r ol l e r on t he gr ound s t a t i o n. T h e M A V 128 R 3 O n b oar d C om p u t e r I t w a s e vi de nt f r om t he be gi nni ng t ha t s om e s or t o f e m be dde d pr oc e s s or w a s r e qui r e d onboa r d t he a i r c r a f t T he r e w a s i ni t i a l l y s om e t hought gi ve n t o us i ng a n of f t he s he l f P ow e r P C e m be dde d s ys t e m r unni ng uc L i nu x. S uc h a de vi c e c oul d be i nt e r f a c e d t o i ne r t i a l s e ns or s a nd a G P S a nd i t w a s pow e r f ul e n ough t ha t i t c oul d e a s i l y r un a c ont r ol l e r onboa r d a nd pos s i bl y e ve n pa r t o f t he vi s i on s ys t e m H ow e ve r t he s i z e w e i gh t a nd pow e r r e qui r e m e nt s o f t hi s de vi c e l e d us t o dr op t hi s a s a s ol ut i on, a s di d t he f a c t t ha t i t s ve r s i on of uc L i nux r e qui r e d ove r a m i n ut e t o boot T he r e w a s s t i l l t he de s i r e t o bui l d a s ys t e m t ha t c oul d ha ndl e t he G P S a nd I M U s ys t e m s f or now but e ve n t ua l l y be F i gur e 2 3 : M H X 2400 R F T r a ns c e i ve r

PAGE 23

12 c a pa bl e of s uppor t i ng t he hor i z on t r a c ki ng s ys t e m onboa r d. T he r e f or e a D S P s ys t e m w a s c ons i de r e d. H ow e ve r w e onl y ha d a f e w m ont hs t o de ve l op a f l i ght s ys t e m c a pa bl e of us i ng t he vi s i on a nd i ne r t i a l s e ns or s t o ke e p t he M A V s t a bl e a nd us i ng G P S t o na vi ga t e T he r e f or e i t w a s de c i de d t o j us t f oc us on t he I M U a nd G P S f or t he onboa r d s ys t e m s T he vi s i on p r oc e s s i ng s ys t e m s w oul d r e m a i n on t he g r ound s t a t i on c om put e r f or now F or t hi s s i t ua t i on, a no r m a l e m b e dde d m i c r oc ont r ol l e r c oul d be us e d. A n A t m e l A V R M e ga 128 m i c r oc ont r ol l e r w a s s e l e c t e d, a de vi c e w e ha d us e d i n pa s t pr oj e c t s T hi s w a s t he m os t pow e r f ul pr oc e s s or i n t he A V R f a m i l y a t t he t i m e c a pa bl e of r unni ng a t up t o 16 M H z a nd gi v i ng up t o 16 M I P S o f t h r oughput I t c ont a i ns 128K of F l a s h M e m or y f or P r og r a m s a nd 4K of R A M f or da t a [ 14] B e s i de s i t s pe r f or m a nc e a nd t he f a c t t ha t w e w e r e a l r e a dy f a m i l i a r w i t h t hi s de vi c e t he M e ga 128 w a s w e l l s ui t e d t o t he p r oj e c t f o r t he f ol l ow i ng f e a t ur e s I t ha d up t o e i ght P W M out put s w hi c h w e r e ne c e s s a r y f or c ont r ol l i ng s e r vos a s w e l l a s a n e i ght c ha nne l 10 bi t A D C f or r e a di ng a na l og s e ns or s W hi l e m os t m i c r oc ont r ol l e r s ha ve onl y one U A R T t he M e ga 128 ha s t w o, a l l ow i ng c ont r ol o f t w o s e pa r a t e s e r i a l d e vi c e s A c c om pi l e r w a s a va i l a bl e f or t hi s de vi c e a l l ow i ng us t o a voi d a s s e m bl y a nd t he r e f or e m or e qui c kl y de ve l op t he c ode f or t he f l i ght s ys t e m F i g ur e 2 4 : M A V 128 R 1, R 2

PAGE 24

13 D e ve l opm e nt of a n onboa r d f l i ght c om put e r be ga n i n t he s p r i ng o f 2003 D ue t o t he f a c t t ha t i t w a s i nt e nde d f o r s m a l l a i r c r a f t a nd bui l t a r ound a M e ga 128, t he f l i ght c om put e r w a s de s i gna t e d a s t he M A V 128. T he f i r s t t w o ve r s i ons s how n i n F i gur e 2 4 w e r e e ngi ne e r i ng p r ot ot ype s a nd t a r ge t e d f or ge ne r a l r e s e a r c h i nt o M A V t e c hnol ogi e s a s w e de t e r m i ne d t he r e qui r e m e nt s of a gi l e a ut onom ous f l i ght T he M A V 128 R 3 w a s s pe c i f i c a l l y t a r ge t e d f or t he A V C A A F s ys t e m de m ons t r a t e d i n J ul y ( F i gur e 2 5) W e w e r e uns ur e a s t o a s t o how w e l l w e w oul d be a bl e t o i s ol a t e t he f l i ght s ys t e m f r om t he m ot or s a nd ke e p t he e l e c t r i c a l noi s e t he y ge ne r a t e d f r o m i nt e r f e r i ng w i t h t he s ys t e m T he r e f or e a g r ound a nd pow e r pl a ne w e r e us e d w i t h a l l of t he f l i ght c om put e r s U s i ng t he s e pl a ne s a l s o a l l ow e d us t o c ut dow n on t he num be r o f t r a c e s t h a t w e ne e de d t o l a yout on t he de s i gn, a nd a s a r e s ul t m i ni m i z e t he s i z e of t he c i r c ui t boa r d W e a l s o s a ve d s pa c e by us i ng s m a l l e l e c t r i c a l t r a c e w i dt hs a nd c l e a r a nc e r e qui r e m e nt s A s i de f r om pow e r s i gna l t r a c e s w hos e hi gh c u r r e nt s r e qui r e d w i de r pa t hs a l l t r a c e w i dt hs a nd c l e a r a nc e s w e r e a t m os t 8 m i l s ( 008 i n. ) a l l ow i ng us t o ke e p t he de vi c e s c l os e t oge t he r E ve n t hi s s pe c i f i c a t i on pr ove d t o be t oo l a r ge a nd w e ha ve s i nc e m ove d t o 6 m i l s c l e a r a nc e a nd t r a c e w i dt hs D ue t o w e i ght a nd s i z e c ons i de r a t i ons s ur f a c e m o unt de vi c e s a nd ot he r s m a l l c om pone nt s w e r e us e d w he ne ve r pos s i bl e W hi l e t hi s w a s e a s y t o a c c om pl i s h i n F i gur e 2 5 : M A V 128 R 3

PAGE 25

14 pr oc ur i ng r e s i s t or s c a pa c i t or s a nd e ve n m os t of t h e I C s t he m a t t e r o f t he c onne c t or s w a s m or e di f f i c ul t T o c ons e r ve s pa c e w e de s i r e d s m a l l m i c r o he a de r s of 05 p i t c h S om e c onne c t or s ha d t o r e m a i n a t t he s t a nda r d he a de r s i z e of 10 pi t c h due t o t he f a c t t ha t t he s m a l l e r c onne c t or s c oul d not ha ndl e t he c u r r e nt f l ow i ng t h r ough t he c a bl e s T he r e w a s a l s o t he pos s i bi l i t y t ha t e xt e r na l c om pone nt s s uc h a s t he G P S I M U a nd R F T r a ns c e i ve r w oul d ne e d t o be c onne c t e d i n t he f i e l d, a nd t h i s w a s m or e di f f i c ul t w i t h s m a l l e r he a de r s W e w e r e a l s o uns ur e i f t he y c oul d ha ndl e t he vi b r a t i on o f t he p l a ne T he r e f or e w e c ont i nue d t o us e s t a nda r d he a de r s f o r pow e r a nd c onne c t i ons t o a l l t he e xt e r na l c om pone nt s of t he f l i ght s ys t e m ( I M U G P S a nd t r a ns c e i ve r ) T he por t s on t he M e ga 128 i t s e l f us e d t he m i c r o he a de r s s i gni f i c a nt l y r e duc i ng t he boa r d s i z e of t he M A V 128. S m a l l c a bl e s c oul d t he n be m a de t o c o n ne c t t o t he s e he a de r s gi vi ng us a c c e s s t o t he i nt e r na l pe r i phe r a l s o f t he M e ga 128. T hi s m e t hod w a s us e d f or t he pr ogr a m m i ng por t bu t t he c a bl e s w e r e f ound t o not be ve r y s e c ur e a nd t he r e f or e w e r e not s ui t a bl e f or f l i ght A l a t e r de ve l opm e nt w a s t ha t o f da ught e r boa r ds w hi c h us e d t he m i c r o he a de r por t s t o pr ov i de a dde d f unc t i ona l i t y t o t he M A V 128. S i nc e m ul t i pl e he a de r s w e r e us e d, t he da ught e r boa r ds w e r e f a r m or e s e c ur e t he n i ndi vi dua l c a bl e s T w o s uc h boa r ds w e r e de ve l ope d ( F i gur e 2 6) T he f i r s t w a s t o a l l ow t he M A V 128 t o s t or e t he t e l e m e t r y i n onboa r d e xt e r na l f l a s h f or l a t e r r e t r i e va l T he da ught e r boa r d c onne c t e d f our s e pa r a t e ba nks of 64 M bi t A t m e l D a t a F l a s h t o t he S P I por t of t he M e ga 128. T he m a j o r f oc us o f t he pr oj e c t s hi f t e d t ow a r ds ge t t i ng t e l e m e t r y t o t he gr ound f o r de ve l opi ng a c ont r ol l e r a nd t hus t he da t a l oggi ng c a pa bi l i t y w a s ne ve r us e d. T he s e c ond da ught e r boa r d a l l ow e d t he M e ga 128 t o c ont r ol up t o f our s e r vos a nd i nt e r f a c e d t o t he i r i nt e r na l pot e nt i om e t e r s

PAGE 26

15 s o t ha t t he M A V 128 c oul d r e c or d t he a c t ua l s e r vo pos i t i ons T hi s a l l ow e d us t o s uppor t t he M A V us e d f or t he r e s e a r c h, w hi c h i nc l ude d t h r e e a c t ua l s e r vos a s w e l l a s t he dr i ve m ot o r w hos e s pe e d w a s c ont r ol l e d i n a s i m i l a r m a nne r t o a s e r vo. T he boa r d a l s o i nc l ude d t he M ot or ol a M P X 4115 A a pr e s s ur e s e ns or f or m e a s ur i ng a l t i t u de w hi c h c oul d be r e a d by t he M e ga 128 s i nt e r na l A D C I n de s i gni ng t he M A V 128, w e a l s o ha d t o c ons i de r t he pow e r r e qui r e m e nt s of t hi s boa r d. T he pr i nc i pl e pow e r s our c e f or M A V s i s c ur r e nt l y L i t hi um P ol ym e r ba t t e r i e s T he l a r ge t hr e e c e l l ba t t e r i e s pr ovi de a r ound 12 vol t s w hi l e t w o c e l l s pr ovi de 7. 4 vol t s T he m a j or l oa d on t he ba t t e r y how e ve r i s t he dr i ve m o t or T he m ot o r i n our s ys t e m r e qui r e s a t hr e e c e l l ba t t e r y T hi s r e s ul t s i n a f l i gh t t i m e of a r ound 20 t o 30 m i nut e s S i nc e t he c ur r e nt dr a w of t he onboa r d e l e c t r on i c s i s i ns i gni f i c a nt c om pa r e d t o t he f l i ght m ot or w e di d not ha ve t o w or r y a bout w he t he r t he e l e c t r on i c s w oul d be t he c r i t i c a l f a c t o r i n t he dur a t i on of a f l i ght T he r e f o r e onc e a ga i n, w e i ght w a s t he pr i m a r y f a c t or A l l o f t he e l e c t r oni c s on t he M A V 128 ope r a t e d a t 5V a nd s o t he ba t t e r y vol t a ge ha d t o be r e gul a t e d. I f w e w e r e t o us e t he s a m e t hr e e c e l l L i P ol y ba t t e r y a s t he m ot or t he n t he di f f e r e nc e be t w e e n t he i nput a nd out pu t vol t a ge s w oul d ha ve be e n s i gni f i c a nt e nough t ha t t he t he r m a l di s s i pa t i on of t he r e gul a t or c oul d pot e n t i a l l y r e s ul t i n t he de vi c e ove r he a t i ng a nd s hut t i ng dow n. F i gur e 2 6 : M A V 128 R 3 D a ught e r B oa r ds

PAGE 27

16 W hi l e t he pr obl e m of t he r m a l di s s i pa t i on c oul d ha ve be e n c om pl e t e l y e l i m i na t e d by us i ng a S w i t c hi ng R e gul a t or s uc h a de vi c e r e q ui r e d e xt e r na l c om pone nt s s uc h a s l a r ge i nduc t or s a nd c a pa c i t or s a nd t he r e f o r e i nc r e a s e d t he s i z e of t he boa r d. W e ha d t o ke e p t he boa r d a s s m a l l a s pos s i bl e a nd s o i ns t e a d w e de c i de d t o pow e r t he e l e c t r oni c s on a s e pa r a t e t w o c e l l ba t t e r y i ns t e a d, a l l ow i ng us t o us e a s t a nda r d vol t a ge r e gul a t or i n a s ur f a c e m ount pa c ka ge e ve n t hough t hi s r e s ul t e d i n a s ys t e m w i t h a n e f f i c i e nc y of onl y 67% I ns t e a d of a L i ne a r R e gul a t or w e u s e d a L D O V ol t a ge R e gul a t or w hi c h c oul d ope r a t e e ve n w he n t he ba t t e r y i nput vol t a ge dr opp e d t o l e s s t he n 6V T hi s a l l ow e d us t o us e t he s a m e e l e c t r oni c s ba t t e r y f or m ul t i p l e f l i ght s a nd onl y ha ve t o r e pl a c e t he m ot or ba t t e r y t o ge t t he pl a ne ba c k i n t he a i r T he c hoi c e f or t he pr i m a r y vol t a ge r e gul a t or o f t he s ys t e m w a s a N a t i ona l S e m i c onduc t or s L M 2940C T h i s L D O r e gul a t or c oul d t a ke t he ba t t e r y i nput a nd pr oduc e up t o 1A of c ur r e nt a t 5 V F l i gh t S ys t e m an d I n t e gr at i on I n t he e nd a l l of t he i ndi vi dua l c om pone nt s ha d t o be c om bi ne d t o w o r k t oge t he r a nd a ny i s s ue s i r one d out s o t ha t a c ont r ol l e r c oul d s uc c e s s f ul l y ke e p t he a i r pl a ne i n t he a i r a nd na vi ga t e p r ope r l y. A m a j o r f oc us w a s on t h e da t a a nd pow e r i nt e r f a c e be t w e e n t he M A V 128, 3D M G I M U S w i f t A 2 G P S R a di o, a n d t he M H X 2400 R F T r a ns c e i ve r A m a j or i s s ue w a s t he f a c t t ha t a l l t h r e e de vi c e s c onne c t i ng t o t he M A V 128 r e qui r e d a U A R T but t he M e ga 128 onl y ha d t w o U A R T S a v a i l a bl e V a r yi ng s our c e s of pow e r w e r e a l s o a n i s s ue T he 3D M G r e qui r e d a t l e a s t 7 V a s i t us e s i t s ow n r e gul a t or s T he M H X 2400 r e qui r e d 5V t o ope r a t e w hi l e t he S w i f t A 2 c oul d onl y ha ndl e 3 3V I n t he e nd t he R F T r a ns c e i ve r w a s us e d on U A R T 0 a nd t he 3D M G on U A R T 1. C ode w a s w r i t t e n f or t he M e ga 128 t o a l l ow one o f t he t i m e r s out put c o m pa r e a nd i nput c a pt ur e f unc t i ons t o ope r a t e a s a s of t w a r e U A R T T hi s ps e udo c om m u ni c a t i ons por t c oul d on l y ope r a t e a t l ow

PAGE 28

1 7 s pe e ds of 4800 bi t s pe r s e c ond or l e s s but t hi s w a s not a n i s s ue f or t he G P S w hi c h c oul d onl y upda t e i t s pos i t i on e ve r y s e c ond. A s t a nda r d c onne c t or w a s s pe c i f i e d f or a l l f l i gh t ha r dw a r e us i ng a U A R T t o m a ke i t e a s y t o de bug t he s ys t e m P i n 1 c ont a i ne d da t a f r om t he M A V 128, a nd P i n 2 da t a t o t he pr oc e s s or P ow e r a nd gr ound w e r e de s i gne d t o be on pi ns 3 a nd 4, r e s pe c t i ve l y. W e ha c k e d t he 3D M G s c a bl e s o t ha t i t c oul d c on ne c t t o t hi s s t a nda r d U A R T por t I nt e r f a c e boa r ds w e r e de ve l ope d f or t he M H X 240 0 a nd S w i f t A 2 G P S s o t ha t t he y c oul d a l s o i nt e r f a c e t o t he por t T he S w i f t A 2 on l y r e qui r e d 150 m A a t 3. 3V T he r e f or e 5V f r om t he M A V 128 R e gul a t or w a s s e nt t hr ough t he U A R T 2 por t a nd a L M 3940I M P 3 3 vol t a ge r e gul a t or f r om N a t i ona l S e m i c onduc t or w a s us e d on t he i nt e r f a c e boa r d t o c onve r t i t t o t he G P S ope r a t i ng vol t a ge T he L M 39 40 w a s c hos e n be c a us e i t w a s i n t he s a m e pa c ka ge a s t he M A V 128 vol t a ge r e gul a t or a nd w a s s pe c i f i c a l l y m e a nt f or c onve r t i ng 5V t o 3. 3V T he i nt e r f a c e boa r d a l s o ha d t he s pe c i a l c onne c t or ne c e s s a r y t o m a t e w i t h t he r i bbon c a bl e c om i ng out of t he S w i f t A 2. A s t h i s c a bl e w a s ve r y s m a l l a nd f l a t w e de c i de d i t w a s t oo di f f i c ul t t o ha c k, a nd l e f t t hi s c a bl e i n t he s ys t e m W hi l e i t w a s f e l t t ha t t he M A V 128 r e gul a t or c oul d ha ndl e t he c ur r e nt r e qui r e m e nt s of t he S w i f t A 2 i n a ddi t i on t o i t s ow n s ys t e m s t hi s w a s not t he c a s e w i t h t he M H X 2400. T he R F T r a ns c e i ve r c oul d c ons u m e m or e t he n 550 m A ove r ha l f of t he c ur r e nt t ha t c oul d be s our c e d by t he L M 2940. T he r e f or e a not he r s uc h r e gul a t or w a s us e d on t he M H X 2400 I nt e r f a c e B oa r d T he M A V 128 t he r e f o r e s our c e d t he r a w ba t t e r y c ur r e nt t o t he I M U a nd R F T r a ns c e i ve r a nd pr ov i de u p t o 1A a t 5V t o t he pr oc e s s or s ys t e m s a nd t he G P S R a di o.

PAGE 29

18 W i t h t he c onne c t i ons be t w e e n t he m a i n c om pone nt s of t he f l i ght s ys t e m f i na l i z e d, t he c ode ba s e ( w r i t t e n e nt i r e l y i n C ) w a s de ve l ope d t o a l l ow t he M e ga 128 t o i n t e r f a c e t o bot h t he G P S a nd I M U a nd s e nd t hi s da t a t o t he gr ound i n a b i na r y f or m a t a s w e l l a s r e c e i ve c om m a nds f r om t he gr ound s t a t i on a nd a dj us t t he s e r vo va l ue s T he m a i n pr ogr a m w a s c a l l e d da t a m a v. A s i ngl e gl oba l s t r uc t ur e c a l l e d m a v_da t a w a s de ve l ope d t ha t c ont a i ne d a l l o f t he da t a t h a t w a s t o be r e c or de d f r om t he s e ns or s a nd s e nt t o t he gr ound s t a t i on. T hi s a l l ow e d a l l of t he va r i ous s of t w a r e m odul e s t o a c c e s s one l oc a t i on f or m a ni pul a t i ng a nd t r a ns f e r r i ng da t a a nd m a ke m a na ge m e nt of t he c ode ba s e f a r e a s i e r W e de ve l ope d a s t a nda r d i nt e r f a c e f or i nt e r f a c i ng t o t he U A R T w hi c h w a s t he n us e d f or a l l t hr e e c om m uni c a t i ons po r t s i n t he s ys t e m ( i nc l u di ng t he s of t w a r e U A R T ) T he I M U G P S a nd R F T r a ns c e i ve r s of t w a r e m odul e s a l l w e nt t hr ough t hi s i nt e r f a c e t o t r a ns f e r da t a C ode w a s a l s o w r i t t e n t o r e a d t he s e r vo c om m a nds f r om t he R C s ys t e m a nd i f ne c e s s a r y, us e t he s e c om m a nds t o c ont r ol t he s e r v o m ot or s of t he a i r pl a ne H ow e ve r t hi s f unc t i on w a s not i ni t i a l l y us e d f o r s a f e t y r e a s ons a s i t w a s c ons i de r e d ne c e s s a r y t o c om pl e t e l y bypa s s t he f l i ght ha r dw a r e a t t h i s s t a ge of de ve l opm e nt T he e nd r e s ul t w a s t he a r c hi t e c t ur e s e e n i n F i gur e 2 7. A f t e r t he f l i ght s ys t e m a nd c ode ha d be e n t ho r oug hl y t e s t e d i n t he l a b, w e be ga n f l i ght t e s t s T he M A V 128, 3D M G a nd S w i f t A 2 G P S w e r e m ount e d i nt o a 24 M A V A s t he e f f or t t o de ve l op t he a bi l i t y t o s t or e t e l e m e t r y onboa r d ha d be e n a ba ndone d, t he M H X 2400 w a s a l s o us e d t o s e nd t he I ne r t i a l a nd G P S da t a t o t he gr ound T he t ot a l w e i ght of t hi s s ys t e m i nc l udi ng t he ba t t e r y a nd D a t a l i nk a nt e nna w a s 212 gr a m s T he f l i ght s ys t e m s e nt t e l e m e t r y t o t he gr ound s t a t i on t hr ough t he R F t r a ns c e i ve r A c ont r ol l e r on t he gr ound s ys t e m w a s de ve l ope d t o us e t hi s da t a t o s e nd s e r vo c om m a nds ba c k t o t he

PAGE 30

19 Figure 2-7: Revision 3 Flight System Software Architecture

PAGE 31

20 M A V 128, w hi c h c ont r ol l e d t he s e r vos W hi l e t he c ont r ol l e r w a s be i ng de ve l ope d, t he gr ound s t a t i on pa s s e d t hr ough S e r vo C om m a nds f r om a hum a n pi l ot a l l ow i ng f or ope n l oopi ng t e s t i ng E ve n w i t h ope n l oop c ont r ol w e e xpe r i e nc e d pr ob l e m s T hough s om e of t h i s w a s due t o de l a ys i n c ont r ol l i ng t he a i r c r a f t t h r ough t he gr ound s t a t i on a not he r f a c t or w a s t he ove r a l l w e i ght of t he f l i gh t s ys t e m I t w a s a t t he v e r y e dge of t he pa yl oa d w e i ght t ha t t he 2 4 M A V c oul d ha ndl e a nd a s s uc h, t he a i r pl a ne w a s e xt r e m e l y di f f i c ul t t o c ont r ol T he r e f or e s o t ha t w e c oul d qui c kl y s t a r t w or k i ng on a c ont r ol l e r a 30 w i ngs pa n M A V ( s e e F i gur e 2 8 ) w a s de ve l ope d f o r t e s t i ng t he pr ot ot ype f l i ght s ys t e m O nc e w e ha d s uc c e s s f ul l y de m ons t r a t e d a ut onom ous f l i ght w e c oul d be gi n t o f oc us on de c r e a s i ng t he w e i ght of t he f l i ght s ys t e m I t w a s a t t hi s po i nt t ha t t he m e t hod of s e ndi ng c om m a nds t o t he s e r vos w a s a l s o c ha nge d, t o e l i m i na t e t he de l a ys i n c ont r ol l i ng t he a i r pl a ne T he M A V 128 j us t s e nd da t a t o t he gr ound w hi l e t he C ont r o l l e r on t he G r oun d S t a t i on s e nt S e r vo C om m a nds t hr ough t he R C S ys t e m T h i s m odi f i c a t i on a l ong w i t h us i n g t he 30 M A V pl a t f o r m e l i m i na t e d t he c ont r ol l a bi l i t y pr obl e m s a nd ope n l oop t e s t i ng be ga n. I n e r t i a l T e l e m e t r y a nd G P S da t a w e r e bot h s e nt ba c k a t a bout 30 H z w hi c h w a s f a s t e nough f or t he C ont r ol s G r oup l e d by P r o f e s s or R i c k L i nd t o be gi n w or k on de ve l opi ng a c ont r ol l e r F i gur e 2 8 : 30" M A V

PAGE 32

21 M or e i s s ue s a r os e a s ope n l oop t e s t i ng c ont i nue d. T he 3D M G w a s s e t t o out pu t bot h i ns t a nt a ne ous a c c e l e r a t i on r a t e s a nd t he E u l e r A ngl e s t ha t de t e r m i ne t he or i e nt a t i on of t he pl a ne T he da t a c om i ng ba c k f r o m t he I M U w a s i nc r e di bl y noi s y. I t w a s i ni t i a l l y t hought t ha t e a r l i e r c r a s he s t ha t ha d r e s ul t e d f r om t he c ont r ol l a bi l i t y i s s ue s ha d da m a ge d t he 3D M G but w he n i t w a s r e pl a c e d w i t h a s e c ond uni t t he s a m e i s s ue s oc c ur r e d. E f f or t s a t f i l t e r i ng w e r e uns uc c e s s f ul i n c l e a ni ng u p t he da t a t o t he poi nt t ha t a c ont r ol l e r c oul d ke e p t he a i r c r a f t s t a bl y f l yi ng. T he r e w a s a p os s i bi l i t y t ha t t he m ount i ng of t he 3D M G m i ght be t he i s s ue A l l o f t he f l i ght s ys t e m c om pone nt s w e r e j us t pl a c e d i ns i de t he f us e l a ge w i t h f oa m ke e pi ng t he m f r om j os t l i n g a r ound i n t he m i dd l e of a f l i ght W e t hought t ha t t h i s m i ght be t oo uns t a bl e f or t he 3D M G s o w oo d i ns e r t s w e r e us e d t o phys i c a l l y m ount t he 3D M G t o t he a i r pl a ne H ow e ve r t he r e w a s s t i l l s i gni f i c a nt noi s e i n t he 3D M G da t a s uc h t ha t t he s e ns or w a s unus a bl e T hi s t r e nd c ont i nue d f or t he ne xt f e w m on t hs a s w e a ppr oa c he d t he J ul y 24t h de m o. A l l e f f or t s t o de ve l op a c ont r ol l e r us i ng t he 3D M G f or i ne r t i a l c ont r ol p r ove d uns uc c e s s f ul S our c e s f or t hi s noi s e w a s t hought t o i nc l ude E M F e m i s s i ons f r om t he da t a a nd vi de o a nt e nna s or pos s i bl y f r om t he d r i ve m ot or of t he M A V i t s e l f T he l a t t e r s e e m e d l i ke l y w he n be nc h t e s t s s how e d pe r f e c t E u l e r A ngl e e s t i m a t i ons f r om t he 3D M G unt i l t he t h r ot t l e w a s t u r ne d on. H ow e ve r m ovi ng t he 3D M G a w a y f r om t he s e pos s i bl e s our c e s di d not he l p t he m a t t e r A na l ys i s by J a s on G r z yw na M uj a hi d A bdul r a hi m a nd m ys e l f f ound t ha t a good de a l of t he pr obl e m w a s t he vi br a t i on of t he a i r pl a ne i t s e l f a nd i t s a f f e c t on t he 3D M G s e ns or s T he M A V w a s put i nt o a r i g t ha t a l l ow e d t he t hr ot t l e t o be s e t t o 100% a nd ye t hol d t he a i r pl a ne i n pl a c e T he pl ot s i n F i gur e 2 9 s how t he 3D M G s e ns or out put s bot h

PAGE 33

22 w i t h t he t h r ot t l e a c t i ve a nd w i t h t he M A V unde r g oi ng vi br a t i on f r om e xt e r na l m ove m e nt w i t h t he m ot o r i na c t i ve T he vi br a t i on a ppe a r s i n bot h s c e na r i os W e t r i e d t o de ve l op s om e f o r m of a dva nc e d f i l t e r t ha t w oul d e na bl e us t o ge t us e f ul da t a out of t he I M U but w i t h l i t t l e s uc c e s s H ow e ve r a d i s c ove r y w a s m a de t ha t t he 3D M G di d ha ve t he a bi l i t y t o pr oduc e f i l t e r e d da t a T he I M U ha d t he a bi l i t y t o us e t he onboa r d gyr os c o pe s t o s t a bi l i z e t he da t a A s t he pl ot i n F i gur e 2 10 s how s doi ng s o i m m e di a t e l y c l e a ne d up t he no i s e on t he E ul e r A ngl e s E ve n w i t h t hi s m odi f i c a t i on t o t he s ys t e m t he r e w e r e s t i l l i s s ue s w i t h noi s e i n t he i ne r t i a l da t a A not he r m a j o r i s s ue di s c ove r e d w a s t ha t t he a c c e l e r om e t e r s on t he 3D M G c oul d onl y ha ndl e g r a vi t a t i ona l f o r c e up t o 2G s H ow e ve r a ny e xt r e m e m a ne uve r i ng F i gur e 2 9 : 3D M G B e nc ht op A na l ys i s

PAGE 34

23 c a us e s t he a i r c r a f t t o e xc e e d t he s e f or c e s a nd t he 3 D M G a c c e l e r om e t e r s t o s a t ur a t e T he s e e xc e s s i ve f or c e s w e r e m os t l y ha ppe ni ng d ur i ng t a ke of f but a l s o oc c ur r e d du r i ng e xt r e m e m a ne uve r s W hi l e t he i ns t a nt a ne ous da t a w oul d r e t ur n t o nom i na l va l ue s t he pr oc e s s by w hi c h t he 3D M G c a l c ul a t e s E ul e r A ng l e s us e s i nt e gr a t i on, a nd t he r e f or e e a c h s uc c e s s i ve c a s e of hi gh g r a vi t a t i ona l f or c e s f ur t he r di s t or t e d t he I M U da t a A n a t t e m pt t o r e s e t t he c a l c ul a t i on pr oc e s s bot h m a nua l l y a nd a t a s e t r a t e pr ove d i ne f f e c t i ve T he pr oc e s s of e xpe r i m e nt i ng w i t h t he 3D M G w a s not he l pe d by t he c ondi t i ons M ul t i pl e a i r pl a ne s w e r e de s t r oye d dur i ng t e s t f l i gh t s r e s ul t i ng i n one 3D M G be i ng de s t r oye d. T he r e s t of t he ha r dw a r e f a r e d s om e w h a t be t t e r t hough t he w e a r a nd t e a r r e s ul t e d i n t he s m a l l c a bl e f or t he S w i f t A 2 de t a c hi ng f r om i t s c onne c t or i ns i de of t he G P S r a di o A s a r e s ul t t he m e t a l c a s e ha d t o be pe e l e d a w a y ( t he r e w a s no w a y t o ope n t he de vi c e ) a nd w i r e s s ol de r e d di r e c t l y t o pi ns on t he i nt e r na l c i r c ui t boa r d of t he G P S T he M A V 128, how e ve r c a m e a w a y f r om a l l c r a s he s uns c a t he d. F i gur e 2 10 : 3D M G B e nc ht op A na l ys i s w i t h D a t a S t a bi l i z a t i on

PAGE 35

24 P r ot ot yp e F l i gh t S ys t e m D e ve l op m e n t : C o n c l u s i on s O ur i na bi l i t y t o ge t a de qua t e da t a out o f t he 3D M G m a de i t i m pos s i bl e t o de ve l op a c ont r ol l e r by t he t i m e of t he de m o a t E gl i n A F B on J ul y 24t h 2004 T hi s w a s t he pr i m a r y r e a s on t ha t w e w e r e una bl e t o de m ons t r a t e a ut onom ous f l i ght I ns t e a d, pa r t s o f t he di f f e r e nt M A V t e c hnol ogi e s c ur r e nt l y be i ng de ve l ope d w e r e s how n, i nc l udi ng a s t a t i c di s pl a y of t he p r ot ot ype f l i ght s ys t e m T he r e w e r e ot he r i s s ue s a s w e l l T he S w i f t A 2 G P S ha d pr ove d t o be i na de qua t e f or our r e qui r e m e nt s T hough i t ha d t he s a m e w e i ght a s t he 3D M G i t s l a r ge f o r m f a c t or m a de i t di f f i c ul t t o ha ndl e e s pe c i a l l y s i nc e i t ha d t o be m ount e d on t he a i r pl a ne s o t ha t i t s i nt e gr a t e d a nt e nna c oul d l oc k ont o G P S s a t e l l i t e s F ur t h e r m o r e t he c a bl e t ha t i t us e d t o c onne c t t o t he ot he r de vi c e s w a s t oo f r a gi l e a nd e xt r e m e l y di f f i c ul t t o r e pl a c e B e s i de s t he I M U pr obl e m s w e a l s o ha d t o c ons i de r t he ove r a l l w e i ght o f t he f l i ght s ys t e m A t a l m os t 250 g r a m s i t w a s s i m pl y t oo he a vy f or a 24 w i ngs pa n M A V t o ha ndl e T hough a gr e a t de a l of t hi s w e i ght w a s t he R F T r a ns c e i ve r a nd a nt e nna w e s t i l l ne e de d t o l ook i nt o de c r e a s i ng t he w e i ght of t he ot he r c om pone nt s a s w e l l W i t h a i r c r a f t a t t hi s s c a l e e ve r y g r a m c ount e d. T he one m a j or be ne f i t t o c om e out of t he P r ot ot yp e F l i ght S ys t e m w a s t he M A V 128. T he F l i ght C om put e r ha d pr ove d t o be m or e t he n a de qua t e i n ha ndl i ng t he m ul t i pl e t a s ks pl a c e d on i t a nd ha d be e n dur a bl e e nough t o s ur vi ve c r a s he s i n m ul t i pl e a i r pl a ne s A s s uc h, i t w a s t he one m a j or c om pone n t f r om t hi s f i r s t s ys t e m t ha t c ont i nue d t o be de ve l ope d, w i t h ne w ve r s i ons r e m a i ni ng a t t h e he a r t of our f ut ur e f l i ght s ys t e m s I nde e d, t he M A V 128 R 3 ha s s i nc e be e n us e d i n ot he r M A V pr oj e c t s a t t he U ni ve r s i t y of F l or i da A n R 3 boa r d w a s us e d i n t he gr ound s t a t i o n t o pr ov i de a n i nt e r f a c e be t w e e n t he G r ound S t a t i on C om put e r a nd a n R C C ont r ol l e r I t a l s o ha s be e n us e d f or s e r vo c ont r ol

PAGE 36

25 a nd t r a ns m i t t i ng t e l e m e t r y i n l a r ge r a i r c r a f t f or f l i ght t e s t s s uppor t i ng M A V r e s e a r c h. T he t e c hnol ogy w a s e ve n m ove d i nt o t he P oc ke t M A V a s m a l l 12 w i ngs pa n M A V t ha t i s t o a l l ow f or a ugm e nt e d c ont r ol of a M i c r o A e r i a l V e hi c l e T he i ni t i a l s ys t e m a c t ua l l y us e d a M A V 128 R 3 w i t h t he a ddi t i on o f a da ught e r boa r d c ont a i ni ng a t w o a xi s a c c e l e r om e t e r a nd t he i nt e gr a t i on o f t he t w o P C B s a nd t he r e m ova l of unne c e s s a r y c om pone nt s f or t he s ys t e m l e d t o t he M A V 128 ( F i gur e 2 11 ) a c i r c ui t boa r d w i t h di m e ns i ons 7 x 7 [ 15 ] F i gur e 2 11 : M A V 128 F l i gh t C ont r ol l e r

PAGE 37

26 C H A P T E R 3 R E V I S I O N 4: F L I G H T S Y S T E M W I T H O N B O A R D G P S W i t h t he J ul y de m o c om pl e t e a t t e nt i on t ur ne d t o t he ne xt de m ons t r a t i on, w hi c h w a s s uppos e d t o oc c ur i n l a t e O c t obe r W e s t i l l ha d t he goa l o f de ve l opi ng a f l i gh t s ys t e m t o e na bl e a ut onom ous f l i ght a nd n a vi ga t i on of a 24 M A V T o a c c om pl i s h t hi s w e ne e de d t o bot h r e pl a c e t he 3D M G a nd d r a m a t i c a l l y r e duc e t he w e i ght of t he f l i gh t s ys t e m I t a ppe a r e d t ha t t he onl y w a y t o c r e a t e a c ont r ol l e r c a pa bl e of s t a bi l i z i ng t he a i r c r a f t dur i ng f l i ght w a s t o de ve l op o ur ow n I M U s ys t e m di r e c t l y i n t e gr a t e d w i t h t he M A V 128. H ow e ve r doi ng s o i n l e s s t he n t hr e e m ont hs w a s n e a r i m pos s i bl e e s pe c i a l l y s i nc e w e a l s o ne e de d t o a l s o f i nd r e pl a c e m e nt s f or t he R F T r a ns c e i ve r a nd t he G P S t o c ut t he w e i ght of t he ove r a l l f l i ght s y s t e m a nd de ve l op t h e R e vi s i on 4 M A V 128 t o s uppor t t he ne w s ys t e m r e qui r e m e nt s A t e m po r a r y s ol ut i on w a s t o us e t he vi s i on gui de d s t a bi l i t y s ys t e m I ns t e a d of ha vi ng a n I M U onboa r d f o r t he O c t obe r de m o, t he c ont r ol l e r us e d a n a ugm e nt e d ve r s i on of t he hor i z on t r a c ki ng s ys t e m t o obt a i n t he i ns t a nt a ne ous r ol l a nd pi t c h a ngl e s T hi s da t a c oul d t he n be us e d t o ke e p t he a i r c r a f t s t a bl e w hi l e f l yi ng T he ne w onboa r d f l i gh t s ys t e m us e d t he M A V 128 R 4 t o obt a i n G P S da t a a nd s e nd i t t o t he gr ound f o r t he c ont r ol l e r t o us e i n na vi ga t i ng t he pl a ne F ur t he r m or e t he g r ound s t a t i on a l s o w a s us e d t o de ve l op ne w vi s i on a l gor i t h m s f o r obj e c t i ve s s uc h a s t a r ge t t e s t i ng. T hi s a l l ow e d t he R e vi s i on 4 s ys t e m t o be us e d a s a t e s t be d f or bot h a dva nc e d vi s i on pr oc e s s i ng s ys t e m s a nd G P S ba s e d na vi ga t i ona l c o nt r ol [ 16 17 ] O nc e t he O c t obe r de m o w a s c om pl e t e w e c oul d t he n s t a r t w or ki ng on de v e l opi ng on t he M A V 128 R 5 w hi c h

PAGE 38

27 w oul d i nc l ude I ne r t i a l S e ns or s a nd be a dr op i n r e p l a c e m e nt f or t he ne w s ys t e m T he R e vi s i on 5 boa r d w oul d a l l ow t he c ont r ol l e r t o be m odi f i e d t o t a ke a dva nt a ge of G P S V i s i on, a nd I N S a nd f ul l y de m ons t r a t e a ut onom ou s f l i ght F l i gh t S ys t e m C om p on e n t s G l ob al P os i t i on i n g S ys t e m R e c e i ve r : F u r u n o G H 80D T he r e w e r e m u l t i pl e r e a s ons f or r e pl a c i ng t he A xi o m S w i f t A 2 G P S R e c e i ve r A m a j or i s s ue w a s i t s w e i ght w hi c h w a s 28 g r a m s A r e l a t e d pr obl e m w a s t he f o r m f a c t or A t 1. 65 by 1. 65 t he S w i f t A 2 t ook up a m a j or p or t i on of t he t op s ur f a c e a r e a o f t he M A V F i na l l y, i t s c onne c t or ha d be e n p r ove n t o b e uns ui t a bl e f or t he s o m e t i m e s r ough c ondi t i ons of M A V f l i gh t t e s t i ng, w i t h t he r e s ul t b e i ng t ha t t he r e w e r e s e ve r a l t r i ps t o t he f i e l d w he r e a l os e c onne c t i on r e s ul t e d i n no na vi ga t i ona l da t a A s e a r c h f or a ne w G P S uni t w a s be gun A not he r M A V gr a nt a t t he U ni ve r s i t y of F l or i da ha d c om e a c r os s a ne w G P S uni t t he G H 80 ( F i gu r e 3 1 ) by a c om pa ny c a l l e d F ur uno, a nd w a s ha vi ng s om e s uc c e s s w i t h i t A f t e r s om e i nve s t i ga t i on, w e s e t t l e d on t he G H 80 f or t he R e vi s i on 4 F l i ght S ys t e m A t 8 x 8 s i z e a nd w i t h a w e i ght o f onl y 12 gr a m s i t s phys i c a l s pe c i f i c a t i ons w e r e w e l l s ui t e d f or ou r ne e ds F ur t he r m o r e t he r e w a s a ve r s i on a va i l a bl e w i t h 10 s m a l l pi ns pr ot r udi ng f r om t he bot t o m of G H 80 t ha t c oul d be us e d f or i nt e r f a c i ng [ 18] A s i m pl e c i r c ui t w a s de s i gne d t ha t a l l ow e d t he G P S t o be di r e c t l y s ol de r e d t o a P C B a nd t he n c onne c t e d t o t he M E G A 128 t h r ough t he s t a nda r d 4 P i n U A R T H e a de r T hi s c om pl e t e l y e l i m i na t e d t he c onne c t or p r obl e m s w e ha d be e n e x pe r i e nc i ng w i t h t he S w i f t A 2. F i gur e 3 1: G H 80D

PAGE 39

28 T he pr ot oc ol us e d t o i nt e r f a c e t o t he G H 80 w a s a m a nuf a c t ur e r de f i ne d b i na r y s pe c i f i c a t i on, but w a s s i m i l a r e nough t o S i R F t ha t i t w a s qui t e e a s y t o r e w r i t e t he G P S i nt e r f a c e c ode f r o m t he S w i f t A 2 t o w or k w i t h t he F ur uno G P S i ns t e a d. O ne i s s ue t ha t w a s di s c ove r e d w a s t ha t t he G H 80D i gno r e d a ny c om m a nds s e nt t o i t f or t he f i r s t f e w s e c onds a f t e r pow e r up. U nf or t una t e l y t h i s i nf or m a t i on w a s not i nc l ude d i n t he pr e l i m i na r y da t a s he e t a nd a s s uc h i t t ook a w hi l e t o de t e r m i ne w hy t he G P S uni t s w e r e i gnor i ng t he c om m a nds be i ng s e nt t ha t s e t w h i c h d a t a f or m a t t he de vi c e s houl d s e nd. O nc e t hi s i s s ue w a s r e s ol ve d by c ont i nuous l y s e ndi ng t he c om m a nds unt i l t he r e w a s a c ha nge i n w ha t da t a w a s g s e nt t he c ode ne c e s s a r y t o i nt e r f a c e t he G H 80D t o a M A V 128 f l i ght c om put e r w a s c om pl e t e R F T r an s c e i ve r : A e r oc om m A C 4490 500 T he M i c r oha r d M H X 2500 ha d be e n s e l e c t e d pr i m a r i l y be c a us e of t he i r a dve r t i s e d a bi l i t y t o t r a ns m i t a t up t o 115. 2 kbps a t 20 m i l e s T hough t he y ope r a t e d a de qua t e l y, t he e xc e s s i v e w e i ght w a s a c r i t i c a l i s s ue H ow e ve r a f t e r a n e xha us t i v e s e a r c h not hi ng e l s e c oul d be f ound t ha t m e t t he r e qui r e m e nt s a nd s o t he M i c r oha r d T r a ns c e i ve r s w e r e c hos e n f or t he R e vi s i on 3 F l i ght S ys t e m i n A p r i l 2 003. I t t ur ns ou t t ha t w e s houl d ha ve c ont i nue d l ooki ng f or vi a bl e a l t e r na t i ve s O ne m ont h l a t e r a c om pa ny c a l l e d A e r oc om m s t a r t e d publ i c i z i ng i t s ne w 900 M H z R F T r a ns c e i ve r t he A C 4490 ( F i gu r e 3 2 ) w hi c h c oul d a l s o c om m uni c a t e a t 115 2K T he de vi c e c a m e i n a f e w di f f e r e nt ve r s i ons c a pa bl e of t r a ns m i t t i ng a t di f f e r e nt pow e r l e ve l s T he A C 4490 500, w hi c h w a s t he m os t pow e r f ul h a d a r a nge of 20 m i l e s m o r e t he n a de qua t e f or our ne e ds F u r t he r m or e t he de vi c e w a s 1. 9 x 1 65 a t a w e i ght o f onl y 12 gr a m s a nd t he r e f or e l e s s t he n ha l f t he s i z e a nd 1 / 7 of t he w e i ght o f t he M H X 2400 [ 19]

PAGE 40

29 T he r e w e r e o f c our s e i s s ue s O ne w a s t he f a c t t ha t t he A C 4490 w a s a ve r y ne w de vi c e F ul l y de t a i l e d s pe c i f i c a t i ons on t he de vi c e w e r e not ye t a va i l a bl e a nd s o s om e o f our w or k i n m a ki ng t he t r a ns c e i ve r s f unc t i on i n ou r s ys t e m w a s t r i a l a nd e r r or s i nc e t he ne c e s s a r y i nf or m a t i on w a s not a l w a ys a t ha nd. F u r t he r m or e t he A C 4490 500 w a s not ye t a va i l a bl e W hi l e i t w a s i ni t i a l l y s uppos e d t o be r e a dy i n t he s um m e r of 2003, i t e nde d up be i ng de l a ye d a nd w e w e r e no t a bl e t o ge t t he f i r s t t w o m odu l e s unt i l l a t e A ugus t A s a r e s ul t w e i ni t i a l l y s t a r t e d us i ng A C 4490 200 R F T r a ns c e i ve r s i ns t e a d. T he y w e r e i de nt i c a l t o t he A C 4490 500 m odul e s e xc e pt t ha t t he y t r a ns m i t a t a l ow e r pow e r w hi c h r e s ul t s i n a m a xi m um r a nge o f f ou r m i l e s F or i ni t i a l t e s t i ng, t h i s w a s not a n i s s ue W e s t a r t e d w i t h de ve l opm e nt ki t s f or t he t r a ns c e i ve r s H ow e ve r w e r a n i n t o s om e pr obl e m s i n t e s t i ng t he de vi c e s w hi c h di d not ope r a t e ve r y w e l l i n t he l a b W e a s s um e d t ha t t he i s s ue s w he n t e s t i ng t he r a di os unde r t he s e c ondi t i ons w a s t he r e s ul t of i nt e r f e r e nc e f r om ot he r de vi c e s i n t he l a b a l s o r unn i ng a t 900 M H z a s w e l l a s f r om t he f a c t t ha t w e w e r e w or ki ng i n c onf i ne d r oom s w i t h w a l l s t ha t pr oba bl y c ont a i ne d s om e m e t a l s hi e l di ng. A s a r e s ul t r a di o r e f l e c t i ons w e r e l i ke l y t o oc c u r I n a ny e ve nt w e f ound w e ha d be t t e r r e s ul t s w he n w e ke pt t he t r a ns c e i ve r s i n s e pa r a t e r oom s a nd m uc h be t t e r pe r f or m a nc e onc e w e m ove d out door s F i gur e 3 2: A C 4490 R F T r a ns c e i ve r

PAGE 41

30 T he r e w a s t he m a t t e r of de t e r m i ni ng how t o s e t up t he m odul e s F o r t una t e l y, A e r oc om m pr e s e t m os t of t he ne c e s s a r y pa r a m e t e r s f or opt i m um pe r f or m a nc e de pe ndi ng on t he s c e na r i o t he A C 4490 s w e r e ope r a t i ng unde r T he r e w a s s t i l l t he i s s ue of s e t t i ng up t he ne t w or k. F i r s t o f a l l one A C 4490 ha d t o be s e t a s a s e r ve r a nd t he ot he r a s t he c l i e nt W e a l s o ha d t o de t e r m i ne w he t he r t he s e r ve r r a di o s houl d be on t he M A V or s t a y w i t h t he gr ound s t a t i on. F u r t he r m or e w e ne e de d t o de t e r m i ne w he t he r t he m a s t e r r a di o s houl d j us t c om m uni c a t e t o t he c l i e nt A C 4490, or b r oa dc a s t t o t he w or l d. I f w e c hos e t he f o r m e r e ve r y t i m e w e c ha nge d c l i e nt s w e w oul d ha ve t o m odi f y t he c on f i gur a t i on of t he s e r ve r r a di o w i t h t he ne w c l i e nt a ddr e s s T he r e w a s a l s o t he m a t t e r o f w he t he r t he r a di os s houl d be s e t t o s t r e a m m ode or a c know l e dge S t r e a m m ode c oul d a c t ua l l y r e s ul t i n a f a s t e r t h r ou ghput but a s a r e s ul t da t a pa c ke t s w e r e s om e t i m e s br oke n up. T hi s l a t e n c y i n t he r e c e i ve d da t a a ppe a r e d t o c a us e i s s u e s i n r e m a i ni ng i n s ync h w i t h t he gr ound s t a t i on, a nd r e s ul t i n c or r upt e d t e l e m e t r y. F u r t he r m o r e t he r a di os w oul d not r e t r y s e ndi ng c or r upt e d pa c ke t s i n t hi s m ode a nd s o e r r or s c oul d oc c ur [ 20] T he r e f or e a c know l e dge m ode w a s s e l e c t e d. W e a l s o a dj us t e d t he ba ud r a t e dow n t o 19 2K a s s e ndi ng s o m uc h da t a a t hi ghe r ba ud r a t e s a ppe a r e d t o r e s ul t i n dr oppe d pa c ke t s a s w e l l B r oa dc a s t m ode a l s o s e e m e d t o r e s ul t i n d r oppe d p a c ke t s a nd s o w e s e t t he s e r ve r r a di o t o c om m un i c a t e w i t h t he s i ngl e c l i e nt W e a l s o not i c e d i s s ue s w he r e t he c l i e nt r a di o oc c a s i ona l l y l os t s ync h w i t h t he s e r ve r W hi l e t he de vi c e us ua l l y r e a c qui r e d t he s ync h, a t ot he r t i m e s t he c l i e nt ne e de d t o be r e s e t A s t h i s w a s i m pos s i bl e t o do w i t h t he M A V A C 4490 w he n i t w a s i n t he a i r w e c hos e t o l oc a t e t he c l i e nt de vi c e i n t he gr ound s t a t i on.

PAGE 42

31 W e ha d be gun t o ga i n a n unde r s t a ndi ng o f t he de vi c e s w he n w e r e c e i ve d our f i r s t A C 4490 5 00s H ow e ve r w e t he n r a n i nt o p r obl e m s be c a us e of t he f a c t t ha t t he hi ghe r pow e r r a di os w e r e not c om pa t i bl e w i t h our de ve l o pm e nt boa r ds T he A C 4490 200 w a s a va i l a bl e i n bot h 5V a nd 3 3V ve r s i ons W e w e r e us i ng t he 5V ve r s i ons t o ha ve l e s s of a vol t a ge dr op w he n c onve r t i ng f r om t he ba t t e r y vol t a ge t o t he R F T r a ns c e i ve r vol t a ge H ow e ve r i t w a s t he n a nnounc e d t ha t t he A C 4490 500 w oul d onl y be a va i l a bl e a t 3 3V T he de ve l opm e nt boa r ds A e r oc om m ha d o r i gi na l l y s e nt us w e r e onl y a n i ni t i a l de s i gn t ha t c oul d not s uppor t 3 3V m odul e s F o r t una t e l y, w e w e r e a bl e t o ob t a i n ne w de ve l opm e nt boa r ds f r om A e r oc om m t ha t ha d m uc h gr e a t e r f unc t i ona l i t y, i nc l udi ng t he a bi l i t y t o i n t e r f a c e t o our ne w R F t r a ns c e i ve r s W e t e s t e d t he r a nge o f t he A C 4490 500s a nd f ound t ha t t he y w e r e r oughl y e qui va l e nt t o t he ol d M H X 2400s S a t i s f i e d, w e now f oc us e d on t he de ve l opm e nt o f t he M A V 128 R 4 a nd t he i nt e gr a t i on of a l l o f t he s e c om pone nt s i nt o t he f l i ght s ys t e m ne c e s s a r y f or t h e O c t obe r de m o. T h e M A V 128 R 4 an d F l i gh t S ys t e m I n t e gr at i on T he M A V 128 R 4, w hos e de s i gn be ga n i n e a r l y J ul y 2003, gr e w out of a de s i r e t o c om bi ne i nt o a s i ngl e de s i gn t he M A V 128 R 3 a nd t he da ught e r boa r d t ha t a dde d s e r vo m ot or s uppor t a nd a n a l t i m e t e r a s w e l l a s i nt e gr a t i ng a ny ne c e s s a r y i nt e r f a c e boa r ds f o r t he ne w G P S a nd R F t r a ns c e i ve r de vi c e s S i nc e t he f i r s t t w o pr ot o t ype s ( t he r e w e r e t hr e e di f f e r e nt ve r s i ons of t he M A V 128 f or t hi s ve r s i on of t he f l i ght s ys t e m R 4A 4B a nd 4C ) w e r e c r e a t e d be f o r e i t ha d be e n f i na l l y de c i de d t o dr op t he 3D M G bot h s t i l l s uppo r t e d t he I M U T he f i na l ve r s i on of t he M A V 128 R 4 r e m ove d 3D M G s uppor t T he r e a f t e r t he R F t r a ns c e i ve r w a s l oc a t e d on U A R T 0, a nd t he G P S on U A R T 1. T hi s ha d t he a dde d be ne f i t o f r e m ovi ng t he ne e d f or t he s of t w a r e ba s e d U A R T 2, a nd a s a r e s ul t t he out put c om pa r e a nd i nput c a pt ur e t i m i ng r e s our c e s f or m e r l y r e s e r ve d f or t he U A R T

PAGE 43

32 be c a m e a va i l a bl e f or ot he r f unc t i ons T he l a t t e r w a s e s pe c i a l l y i m por t a nt s i nc e i t r e ope ne d t he pos s i bi l i t y of r e a di ng s e r vo c om m a nds f r om t he onboa r d R C e qui pm e nt T he da ught e r boa r d ha d be e n c r e a t e d s pe c i f i c a l l y s o t ha t t he s e r vos w oul d not ha ve t o c onne c t t o t he m i c r o he a de r s T he r e f or e t he s e r vo por t s on t he R 4 a l s o us e d s t a nda r d 1 pi t c h he a de r s S e r vo f e e dba c k w a s onc e a ga i n a va i l a bl e us i ng t he M e ga 128 s 10 bi t A D C T he c onve r t e r a l s o s uppor t e d t he M P X 4115A pr e s s ur e s e ns or f or m e a s ur i ng a l t i t ude w hi c h w a s a l s o i nt e g r a t e d i nt o t he R 4. T he pr ogr a m t ha t w e de ve l ope d f or t he M A V 128 R 4 w a s a c t ua l l y r a t he r s i m i l a r t o t ha t of t he R e vi s i on 3 s ys t e m T he 3D M G c ode w a s of c our s e r e m ove d, a nd t he G P S i nt e r f a c e s of t w a r e m odi f i e d t o pr oc e s s t he F ur uno P a c ke t s i ns t e a d of S i R F W i t h t he r e m ova l of t he s of t w a r e U A R T t he M e ga 128 s i nput c a pt ur e 3 ha r dw a r e w a s a va i l a bl e a nd c ode t ha t ha d be e n de ve l ope d e a r l i e r t o r e a d s e r vo c om m a nds f r o m a n R C a i r c r a f t r e c e i ve r w a s r e i ns e r t e d i nt o t he pr og r a m I n a n a t t e m pt t o e l i m i na t e s ync hr oni z a t i on pr obl e m s t ha t s om e t i m e s oc c ur r e d be t w e e n t he M A V a nd t he gr ound c om put e r t he M A V 128 onl y s e nt a pa c ke t of t e l e m e t r y w he n i t r e c e i ve d a r e que s t f r om t he G r ound S t a t i on. H ow e ve r t he s of t w a r e m os t l y r e m a i ne d t h e s a m e a s s e e n i n F i gur e 3 3 M os t of t he c ha nge s t hr ough t he s e r i e s of M A V 12 8 B oa r ds r e s ul t e d f r om de t e r m i ni ng how t o s uppl y t he A C 4490 w i t h pow e r M uc h o f t he s e i s s ue s w e r e due t o t he f a c t t ha t w e ha d a c c e s s onl y t o pr e l i m i na r y da t a T he A C 4490 i nc l ude d bot h 5V a nd 3 3V ve r s i ons W e t he r e f or e de s i gne d t he f i r s t M A V 128 R 4 boa r d t o s uppl y 5V f r om t he m a i n vol t a ge r e gul a t or onboa r d ( F i gur e 3 4) A c onne c t or l oc a t e d on t he bot t om of t he boa r d a l l ow e d t he A C 4490 t o c onne c t d i r e c t l y unde r ne a t h t he M A V 128, a nd c om m uni c a t e di r e c t l y t o t he A T m e ga 128.

PAGE 44

33 Figure 3-3: Rev. 4 Flight Sy stem Software Architecture

PAGE 45

34 T hi s pr oc e dur e w o r ke d r a t he r w e l l f or t he A C 4490 200, a nd t he onboa r d s uppl y w a s a bl e t o pow e r t he A C 4490, t he M A V 128, a nd t he G H 80, w hos e 3. 3V s uppl y w a s de r i ve d f r om t he s a m e s our c e H ow e ve r i t w a s t he n a nnounc e d t ha t onl y 3 3V ve r s i on o f t he A C 4490 500 w e r e a va i l a bl e G i ve n t he f a c t t ha t no pow e r c ons um pt i on da t a w a s a va i l a bl e w e a s s um e d t ha t t r a ns m i t t i ng 5W a t 3. 3V c ons um e d a bout 150 m A W e ha d a l r e a dy s e l e c t e d a vol t a ge r e gul a t or f or t he G P S t o r e pl a c e t he ol d L M 3940, a nd de c i de d t o us e i t f o r t he A C 4490 500 a s w e l l S e l e c t e d f or i t s ve r y s m a l l s i z e o f 3 x 3 m m t he R E G 113N A 3. 3 c oul d s our c e up t o 400 m A A s t he G P S i t s e l f c oul d c ons um e a l m os t 90 m A w e t hought i t be t t e r t o us e s e pa r a t e r e gul a t or s f or t he t w o 3. 3V de vi c e s on t he M A V 12 8 R 4B ( F i gur e 3 4) H ow e ve r w e i m m e di a t e l y ha d pr obl e m s t r a ns m i t t i ng da t a t hr ough t he A C 4490 50 0 w he n us i ng t he M A V 128 R 4B t o s our c e t he pow e r I t onl y w or ke d f o r a m i nut e o r t w o a nd t he n onl y w i t h i nt e r m i t t e nt c om m uni c a t i on T h e I n R a nge S i gna l on t he C l i e nt A C 4490, w hi c h s houl d a s s e r t w he n i t i s s ync he d w i t h a S e r ve r c ons t a nt l y s w i t c he d on a nd o f f W e i m m e di a t e l y s us pe c t e d t ha t t he de vi c e w a s a c t ua l l y c ons um i ng m or e pow e r t he n or i gi na l l y t hought A n i nt e r f a c e boa r d w a s c r e a t e d us i ng t he ol d L M 2940 / L M 3940 s ys t e m f r om t he R e vi s i on 3 S ys t e m t o c onve r t a ba t t e r y s our c e t o 5V a nd t he n t o 3 3V H ow e v e r t hi s s ys t e m ha d t he s a m e pr obl e m s F i gur e 3 4: M A V 128 R 4A R 4B

PAGE 46

35 O ur a t t e nt i on t he n t ur ne d t o w he t he r pl a c i ng t he boa r d unde r ne a t h t he M A V 128 c oul d be r e s ul t i ng i n s om e t ype of e l e c t r i c a l noi s e i nt e r f e r i ng w i t h t he ope r a t i on of t he A C 4490. W e oc c a s i ona l l y s a w i ndi c a t i ons t ha t t hi s c oul d be a f a c t or s uc h a s m ovi ng t he t r a ns c e i ve r a w a y f r om t he M A V 128 r e s ul t i ng i n t h e de vi c e t r a ns m i t t i ng a ga i n. H ow e ve r i n m os t c a s e s t he pos i t i on of t he de vi c e a ppe a r e d t o ha ve no a f f e c t on w he t he r t he R F t r a ns c e i ve r w a s ope r a t i ng, a nd w e de c i de d t ha t a ny i ndi c a t i ons of i nt e r f e r e nc e f r om t he M A V 128 be i ng a f a c t or w a s a c oi nc i de nc e W e w e r e a l s o uns ur e i f m a ybe s om e s i gna l s t ha t w e r e de a l t w i t h on t he de ve l opm e nt boa r d s uc h a s R T S a nd C T S ne e de d t o be i nt e r f a c e d on t he M A V 128 B oa r d a s w e l l f o r t he d e v i c e t o ope r a t e pr ope r l y. H ow e ve r t he A C 4490 200 ha d not de m ons t r a t e d a ny p r obl e m i n t hi s a r e a a nd t he da t a s he e t s a i d t ha t t he y c oul d be l e f t f l oa t i ng i f unus e d. T o ve r i f y w e m a de s ur e a l l o f t he i nput s t o t he A C 4490 ot he r t he n T X D t r a ns m i t da t a w e r e d e a s s e r t e d. D oi ng s o ha d no a f f e c t E ve nt ua l l y, f e e dba c k f r om A e r oc om m a s t o t he a c t ua l pow e r c ons um pt i on o f t he de vi c e de t e r m i ne d t ha t t he pr obl e m ha d a l l a l ong b e e n pr ovi di ng e nough pow e r t o t he A C 4490 500. W he n t r a ns m i t t i ng 100% of t he t i m e t he r a di o c o ns um e d up t o a l m os t 5A T hough t hi s e xpl a i ne d w hy t he R F T r a ns c e i ve r w a s not w or ki ng w he n pow e r e d by t he M A V 128 R 4B i t f a i l e d t o e xpl a i n w hy t he i n t e r f a c e boa r d di d not w or k 1 L M 2 9 4 0 D a t a s h e e t N a t i o n a l S e m i c o n d u c t o r 2 0 0 3 F i gur e 3 5: S O T 223 M a xi m um P ow e r D i s s i pa t i on 1

PAGE 47

36 W e s oon de t e r m i ne d t he c a us e how e ve r w e ha d f a i l e d t o e ns ur e t he c ondi t i ons us e d f or t he L M 2940 / L M 3940 P ow e r S ys t e m on t he i nt e r f a c e boa r d w e r e t he s a m e a s t he L M 2940 / R E G 113 s ys t e m on t he M A V 128 R 4B W e ha d a l w a ys pow e r e d t he f l i ght c om put e r on a t w o c e l l ba t t e r y w i t h a s our c e o f a r ound 7. 4V W he n w e pow e r e d t he i nt e r f a c e boa r d, how e ve r w e ha d us e d a be nc h t op pow e r s uppl y s our c i ng a r ound 12V W hi l e t he L M 2940 c a n ha ndl e a n i nput vol t a ge of ove r 20V t he r e i s a l s o t he m a t t e r of be i ng a bl e t o di s s i pa t e t he t he r m a l e ne r gy t ha t i s g e ne r a t e d by t he di f f e r e nt i a l vol t a ge be t w e e n t he i nput a nd ou t put m ul t i pl i e d by t he c ur r e nt be i ng s our c e d. I n t hi s c a s e t he L M 2940 ne e de d t o di s s i pa t e a bout 3 5W A s s e e n i n F i gur e 3 5, t hi s w a s be yond t he c a pa bi l i t i e s of t he S O T 223 s ur f a c e m ount pa c ka g e w e w e r e us i ng, e s pe c i a l l y s i nc e t he de vi c e i s m e a nt t o di s s i pa t e he a t t hr ough t he g r oun d pl a ne of t he P C B W e w e r e us i ng a pr ot ot ypi ng de vi c e t o qui c kl y c r e a t e t he i nt e r f a c e b oa r ds i ns t e a d of ha vi ng t he m f a br i c a t e d a nd s e nt t o us a f t e r a w e e ks de l a y. T he d r a w ba c k w a s t ha t t he c i r c ui t boa r ds us e d onl y one or t w o l a ye r s w i t h no i nt e r na l pl a ne s A s a r e s ul t t he vol t a ge r e gul a t or ha d l i t t l e a bi l i t y t o di s s i pa t e t he r m a l e ne r gy, a nd t he l o a d w e w e r e pl a c i ng on i t c a us e d t he de vi c e t o s hut dow n t o p r ot e c t i t s e l f a nd on l y t ur n ba c k on w he n i t s i nt e r na l t e m pe r a t ur e ha d r e t ur ne d t o nor m a l l e ve l s O nc e w e us e d a s our c e of 7V f or t he A C 4490 500 i nt e r f a c e boa r d, i t w or ke d pr ope r l y. W i t h t he pr ope r i nf o r m a t i on a nd t he a na l ys i s of t h e pow e r i s s ue s s how n i n T a bl e 3 1, w e de t e r m i ne d t he r e qui r e m e nt s of t he A C 4490, a nd w ha t de vi c e s w e r e r e qui r e d t o pr ovi de pow e r T he r e f or e t he L M 2940 / L M 3940 s ys t e m w a s a l s o us e d i n t he M A V 128 R 4C t he f i na l ve r s i on of t he f l i ght c om put e r i nt e n de d f or t he R e vi s i on 4 f l i ght s ys t e m ( F i gur e 3 6) A s a l w a ys a 2 C e l l L i P ol y ba t t e r y w a s s ue d t o pow e r t he e l e c t r oni c s a nd a s

PAGE 48

37 a r e s ul t t he A C 4490 500 w a s a bl e t o t r a ns m i t w i t hout t he pow e r s ys t e m goi ng i nt o t he r m a l s hut dow n. L a t e r f l i ght t e s t s ve r i f i e d t ha t t he R F T r a ns c e i ve r c oul d s e nd da t a t o t he gr ound, t hough t r a ns m i s s i on s pe e d s w he n s e ndi ng a bout 50 byt e s of da t a w e r e l i m i t e d t o a r o und 10 t o 15H z T a bl e 3 1: A na l ys i s of T he r m a l D i s s i pa t i on I s s ue s i n P ow e r i ng t he R 4 F l i ght S ys t e m V ol t a ge R e gul a t or T he r m a l D i s s i pa t i on P T = I ( V O U T V IN ) A C 4490 200 ( 5V ) C ur r e nt C ons um pt i on 106 m A A C 4490 500 ( 3. 3V ) C ur r e nt C ons um pt i on 492 m A M A V 128R 4 ( 5V ) C ur r e nt C ons um pt i on 10 m A G H 80 ( 3 3V ) C ur r e nt C ons um pt i on 88 m A T w o C e l l B a t t e r y, M A V 128R 4A A C 4490 200, G H 80 L M 2940 5 0V ( 7. 4V 5. 0V ) ( 106 m A + 10 m A + 88 m A ) = 48 96W G H 80 R E G 113N A 3 3V ( 5. 0V 3. 3V ) 88 m A = 1496W T w o C e l l B a t t e r y, M A V 128R 4B A C 4490 500, G H 80 L M 2940 5 0V ( 7. 4V 5. 0V ) ( 492 m A + 10 m A + 88 m A ) = 1 4 16W A C 4490 R E G 113N A 3. 3V ( 5. 0V 3. 3V ) 492 m A = 8364W G P S R E G 113N A 3. 3V ( 5. 0V 3. 3V ) 88 m A = 1496W 12V P ow e r S uppl y A C 4490 500 I nt e r f a c e B oa r d L M 2940 5 0V ( 12V 5 0V ) 492 m A = 3 44W L M 3940 3 3V ( 5. 0V 3. 3V ) 492 m A = 8364W T w o C e l l B a t t e r y, M A V 128R 4C A C 4490 500, G H 80 L M 2940 5 0V ( 7. 4V 5. 0 V ) ( 492 m A + 10 m A + 88 m A ) = 1 4 16W A C 4490 L M 3940 3 3V ( 5. 0V 3. 3V ) 492 m A = 8364W G H 80 R E G 113N A 3 3V ( 5. 0V 3. 3V ) 88 m A = 1496W W i t h t he pow e r i s s ue s c onc e r ni ng t he A C 4490 c om pl e t e our a t t e nt i on t ur ne d t o c onduc t i ng f l i ght t e s t s of t he i nt e gr a t e d f l i ght s ys t e m w hi c h i nc l ude d t he M A V 128 R 4C t he A e r oc om m A C 4490 500, a nd t he F ur uno G H 80D G P S D i r e c t c ont r ol o f t he s e r vos F i gur e 3 6: M A V 128 R 4C W i t h A C 4490

PAGE 49

38 t hr ough t he M A V 128 w a s pos s i bl e i n t w o w a ys w i t h c om m a nds e i t he r s e nt f r om t h e G r ound S t a t i on t hr ough t he A C 4490 da t a l i nk or b y r e a di ng t he s e r vo c om m a nds f r om t he onboa r d R C R e c e i ve r t hr ough t he I nput C a pt ur e on t he M e ga 128. H ow e ve r w e i ns t e a d us e d t he s e t up s how n i n F i gur e 3 8 w i t h t h e s e r vos r e m a i ni ng on t he s t a nda r d R C c ont r ol s ys t e m a nd t he gr ound s t a t i on f l yi ng t he M A V t hr ough t he R C C ont r ol l e r T he r e f or e a l l t he M A V 128 R 4C ha d t o do w a s r e a d i n G P S D a t a a s w e l l a s t a ke r e a di ngs f r om t he a l t i m e t e r t hr ough t he A D C A l l of t hi s da t a w a s t he n s e nt t o t he gr ound s t a t i on t hr ough t he da t a l i nk, w he r e t he c ont r ol l e r c om bi ne d i t w i t h t he pi t c h a nd r ol l a ngl e e s t i m a t i ons f r om t he hor i z on t r a c ki ng s ys t e m t o f l y t he pl a ne G i ve n t he f a c t t ha t w e ha d s i gni f i c a nt l y r e duc e d t h e w e i ght of t he f l i ght s ys t e m i t w a s de c i de d t o t a r ge t a 24 pl a t f or m a ga i n. A M A V t ha t ha d be e n s pe c i f i c a l l y de s i gne d f or t he A V C A A F pr oj e c t ( F i gur e 3 7) w a s t he r e f or e us e d i n our f l i ght t e s t s W e di d not i c e s om e i s s ue s w i t h t he G H 80 G P S r e c e i ve r o nc e w e be ga n f l yi ng A t t i m e s i t t ook up t o a f e w m i nut e s t o l oc k on t o e nough s a t e l l i t e s t o be a bl e t o de t e r m i ne a pos i t i on. I t w a s obvi ous t ha t t he i s s ue w a s e nt i r e l y w i t h t he G P S s e t a nd not t he M A V 128 or A C 4490, be c a us e of t he f a c t t ha t w e w e r e ge t t i ng t e l e m e t r y on t he g r ound a t a l l t i m e s i n t he s e s i t ua t i ons F u r t he r m or e t he t i m e r on t he G P S w a s be i ng r e t r i e ve d a nd s e nt t o t he gr ound a nd i t w a s c ons t a nt l y i nc r e a s i ng, i ndi c a t i ng t ha t t he G P S w a s ope r a t i ona l a nd c om m uni c a t i ng w i t h t he M A V 128. F i gur e 3 7: A V C A A F M A V P l a t f o r m 1 0

PAGE 50

39 S i nc e t he t i m e i t t ook f o r t he G H 80 t o s t a r t p r ovi d i ng a c c ur a t e pos i t i on da t a va r i e d, w e t hought i t w a s pos s i bl e t ha t t he i s s ue w a s t he l o c a t i ons of t he G P S s a t e l l i t e s i n o r bi t a nd t ha t a t c e r t a i n t i m e s t he y w e r e pos i t i one d i n s uc h a m a t t e r t o m a ke i t m or e di f f i c ul t f or t he de vi c e t o obt a i n a l oc k C om m uni c a t i on w i t h E gl i n A F B w he r e t he de vi c e w a s a l s o be i ng us e d, i nc l ude d obs e r va t i ons t ha t t he G H 80D ne e de d a s i z a bl e gr ound pl a ne be ne a t h i t t o i m pr ove t he pe r f o r m a nc e of i t s i nt e g r a t e d a nt e n na H ow e ve r a F u r uno F i e l d A ppl i c a t i on E ngi ne e r c l a i m e d t ha t s uc h a m o di f i c a t i on w a s not ne c e s s a r y, a nd t hough w e di d a dd c oppe r pl a t i ng t o t he ha t c h w he r e t he G P S R e c e i ve r w a s m ount e d, no r e a l i m pr ove m e nt w a s de t e c t e d. I n a ny e ve nt i t o f t e n t ook no m or e t he n a f e w m i nut e s f or t he G P S t o a c qui r e a l oc k, a nd s o t he m a t t e r w a s dr oppe d. A s a p r e c a ut i on, t he c oppe r pl a t i ng r e m a i ne d on t he ha t c h f or a l l of t he R e vi s i o n 4 f l i ght s ys t e m t e s t i ng. F i gur e 3 8: R e vi s i on 4 F l i ght T e s t i ng S e t up

PAGE 51

40 The other issue that was observed as flight testing continued was the sensitivity of the altitude sensor. This device registers the difference in pressure between vacuum and the outside environment, and could range from 15 to 115 kilopascals (kPa). However, this translated to being able to register an altitude in the range of several thousand feet, whereas we were only concerned with regist ering changes on the order of a few hundred feet at most. It was thought that the 10-b it ADC of the Mega128 was capable of giving the system the necessary resolution of a change in a few feet so that the MAV could hold an altitude. However, initial results indicated a resolution of around 40 feet. Given the fact that the Mega128Â’s 10-bit ADC had a resolution of about 5mV when using a 5V reference, we had the ability to register a change of about .1 kPa in pressure. However, this only translated to an accuracy of around 30 feet, as shown in Table 3-2. Furthermore, this was assuming that the ADC in the Mega128 performed with absolute accuracy. In actuality, noise from the processor core could induce errors in the results, and as such, our ability to read changes in altitude was reduced even further unless we were to shut down the core while taking analog readings. Obviously, this was not the solution. Table 3-2: Pressure Sensor Conversion Data ADC Sensitivity V Vref VrefLow High00488 2 1 *10 Pressure Sensor Sensitivity 45.9 mV / kPa ADC Pressure Sensitivity (APS) kPA APS AP S mV kPa mV 1089 5 1 9 45 Altitude vs. Pressure Increase of 1 foot = Decrease of .003559 kPa ADC Altitude Sensitivity (AAS) feet APS kPA APS kPa ft 598 30 1089 003559 1 Instead, it was decided to modify the altimeter circuit. The MPX4115A was initially soldered directly to the board. One was removed from a MAV128 R4C, and a

PAGE 52

41 he a de r i ns t a l l e d t ha t c oul d c onne c t t o a p r ot ot ype b oa r d w hos e pur pos e w a s t o i nc r e a s e t he s e ns i t i vi t y of t he a l t i m e t e r s ys t e m A f e w di f f e r e nt m e t hods w e r e t r i e d O ne o f t he f i r s t a t t e m pt s w a s t o us e ope r a t i ona l a m pl i f i e r s t o m ul t i pl y t he s i gna l B e f or e i t w a s m ul t i pl i e d, a s e t bi a s vol t a ge w a s s ubt r a c t e d f r om t he s e ns or out put s o t ha t t he p r e s s ur e r a nge of i nt e r e s t ( s e a l e ve l t o a t l e a s t a f e w hundr e d f e e t ) e xi s t e d be t w e e n 0 a nd 5V T he t r oubl e w a s t ha t a s t he ga i n f a c t or i nc r e a s e d, r a nge of pr e s s ur e de t e c t a bl e by t h e s ys t e m de c r e a s e d. T he pr e s s ur e a t gr ound l e ve l c oul d va r y de pe ndi ng on l o c a t i on a nd w e a t he r c ondi t i ons m e a ni ng t ha t i f w e s e t t he ga i n f a c t or t oo l a r ge w e r i s ke d be i ng un a bl e t o de t e c t c ha nge s i n a l t i t ude unde r c e r t a i n c ondi t i ons T hough s e ve r a l ga i n f a c t or s w e r e c ons i de r e d a nd us e d dur i ng e xpe r i m e nt s a f a c t or of e i ght w a s u s e d i n a l l f l i ght t e s t i ng, s i nc e i t w a s t he m i ni m um f a c t or t o pr oduc e a t he o r e t i c a l r e s ol ut i on o f 4 f e e t a nd w e w e r e c onc e r ne d t ha t a f u r t he r i nc r e a s e i n ga i n w oul d r e s ul t i n t he A D C s a t ur a t i n g. T o e ns ur e t ha t t he s ys t e m w a s s i m pl e a nd t ha t w e c oul d a c hi e ve t he m a xi m um r a nge pos s i bl e w e r e qui r e d s m a l l s i ngl e s uppl y op e r a t i on a m pl i f i e r s t ha t c oul d p r oduc e a t r ue r a i l t o r a i l out put T hough s uc h de vi c e s w e r e di f f i c ul t t o f i nd w e di d ha ve s uc c e s s w i t h N a t i ona l s L M 324 a nd L i ne a r T e c hnol ogy s L T 1006. T he r e w a s a l s o a n a t t e m pt t o bypa s s t he M e ga 128 A D C a nd us e a T I T L C 3545 14 B i t S e r i a l A D C w hi c h t he n c om m uni c a t e d w i t h t he p r oc e s s or t hr ough a S P I P or t I t w a s hope d t ha t t he f a c t t ha t t hi s de vi c e w a s i s ol a t e d f r om t he P r oc e s s or C or e c om b i ne d w i t h i t s g r e a t e r a c c ur a c y w oul d r e s ul t i n be t t e r a l t i t ude da t a H ow e ve r us i ng t he T L C 3545 di d not pr ovi de a ny obs e r va bl e i nc r e a s e i n a c c ur a c y ove r t he o r i gi na l o pe r a t i on a m pl i f i e r t e s t s ys t e m e ve n w he n t he a l t i m e t e r i nput s i gna l w a s pr e a m pl i f i e d b y t he op a m ps

PAGE 53

42 W e t he r e f or e r e t une d t o w o r ki ng w i t h j us t t he ope r a t i ona l a m pl i f i e r s ys t e m of s ubt r a c t i ng t he bi a s vol t a ge f r om t he s e ns or out put a nd t he n m ul t i pl yi ng t he r e s ul t by e i ght T he i ni t i a l de s i gn us e d a s i ngl e ope r a t i ona l a m pl i f i e r t o ke e p t he s i z e of t he c i r c ui t dow n. H ow e ve r w e f ound t ha t t h i s a ppr oa c h a l l ow e d noi s e on t he bi a s t o ha ve a m a j o r a f f e c t on t he out put A s w e w e r e us i ng a r e s i s t or di vi de r ne t w or k t o ge ne r a t e t he bi a s t hi s w a s t he s i gna l m os t s us c e pt i bl e t o noi s e T he r e f or e w e be ga n us i ng ot he r op a m ps t o buf f e r t he bi a s vol t a ge t he s e ns or out put a nd e ve n t ua l l y e ve n t he a na l og i nput t o t he M e ga 128 A D C a l l i n a t t e m pt t o i s ol a t e t he noi s e f r om t he s e ns or s ys t e m H ow e ve r w e s t i l l ha d a t m os t a pe r f o r m a nc e of a r ound 20 f e e t N e i t he r a m pl i f yi ng t he c i r c ui t by a f a c t or of 8 t hr o ugh t he us e of t he ope r a t i ona l a m pl i f i e r s no r i nc r e a s i ng t he a c c ur a c y of t he A D C by a f a c t or o f 16 ( t hr ough t he T L C 3545) ha d be e n a de qua t e i n i nc r e a s i ng t he pe r f or m a nc e of t he A l t i t ude S e ns or E ve n c om bi ni ng t he t w o s ys t e m s ha d ha d no di s c e r ni bl e a f f e c t I t w a s de c i de d t ha t i f w e w e r e t o a s s um e t ha t t he s e ns or w a s c a pa bl e of pr ov i di ng t he r e s ol ut i on ne c e s s a r y, t he n w e ne e de d t o f ur t he r pr ot e c t t he a l t i t ude s e ns or f r om p os s i bl e i nt e r f e r e nc e T he ne xt s t e p w a s t o i s ol a t e t he pow e r f or t he M e ga 128 f r om t he s e ns or s O ne 5V vol t a ge r e gul a t or ha d t o be t he s our c e f or di gi t a l pow e r a nd gr ound, a not he r f o r a na l og. H ow e ve r s uc h a pow e r s ys t e m r e qui r e d a c om pl e t e r e de s i gn of t he M A V 128. T he r e f or e i t w a s de c i de d t o not s pe nd t i m e m a ki ng t hi s c ha nge f or t he R e vi s i on 4 F l i ght S ys t e m a nd i ns t e a d t hi s w a s i nc or por a t e d i nt o t he m a ny c ha nge s ne c e s s a r y t o t u r n t he M A V 128 i nt o a c om pl e t e I M U s ys t e m f or R e vi s i on 5.

PAGE 54

43 W hi l e t he e xpe r i m e nt s on t he a l t i t ude s e ns or s ys t e m w e r e be i ng c onduc t e d w i t h one M A V 128 R 4C t he ot he r w a s be i ng us e d i n f l i ght t e s t s f or de ve l opi ng t he c ont r ol l e r I ni t i a l a t t e m pt s t o us e t he gr o und s t a t i on ba s e d hor i z on t r a c ki ng s ys t e m t o ke e p t he a i r c r a f t s t a bl e w e r e s uc c e s s f ul [ 21] T he c ont r ol s t e a m t he n m ove d on t o c r e a t i ng t he out e r l oop o f t he c ont r ol l e r s o t ha t t he G P S da t a c o ul d be us e d f or w a ypoi nt na vi ga t i on or f l yi ng be t w e e n di f f e r e nt s e t poi nt s [ 22] A s s e e n i n F i gu r e 3 9 t hi s c ont r ol l e r a l l ow e d t he M A V t o c on t i nuous l y f l y t h r ough a s e r i e s of G P S w a ypoi nt s a nd t he r e f or e de m ons t r a t e d our f i r s t a ut onom ous M A V c a pa bl e of pe r f o r m i ng m i s s i ons F i gur e 3 9: R e v. 4 F l i ght S ys t e m A ut onom ous W a ypoi nt N a vi ga t i on

PAGE 55

44 R e vi s i on 4 F l i g h t S ys t e m C on c l u s i on s W e w e r e a bl e t o de ve l op a f l i gh t s ys t e m t ha t w he n us e d w i t h t he p i t c h a nd r ol l e s t i m a t e d de r i ve d f r om t he g r ound, de m ons t r a t e d a ut onom ous c ont r ol o f t he M A V A l l o f t he m a j or s ys t e m s us e d t he M A V 128 R 4C t he F ur uno G H 80D G P S a nd t he A e r oc om m A C 4490 R F T r a ns c e i ve r ha d p r ove d t o be m or e t he n a de qua t e f o r t he j ob. W e ha d a l s o s uc c e s s f ul l y c ut t he w e i ght of t he ove r a l l f l i gh t s ys t e m by m o r e t he n ha l f f r om 212 gr a m s i n J ul y 2003 t o 86 gr a m s by O c t o be r ( T a bl e 3 3) F ur t he r m o r e t he ha r dw a r e w a s a m a j or s t e p t ow a r ds de ve l opi ng t he R e vi s i on 5 s ys t e m w i t h a n onboa r d I M U T a bl e 3 3: F l i ght S y s t e m s W e i ght D i s t r i but i on ( G r a m s ) S y s t e m P r o t o t y p e M A V 1 2 8 / G P S M A V 1 2 8 / I N S / G P S F l i g h t C o m p u t e r M A V 1 2 8 R 3 1 6 M A V 1 2 8 R 4 1 0 I N S 3 D M G 3 0 M A V 1 2 8 R 5 1 6 G P S S w i f t A 2 2 8 G H 8 0 D 1 2 G H 8 0 D 1 2 D a t a l i n k M H X 2 4 0 0 8 6 A C 4 4 9 0 5 0 0 1 2 A C 4 4 9 0 1 0 0 0 1 2 D a t a l i n k A n t e n n a 2 6 2 6 2 6 B a t t e r y T w o C e l l L i P o l y 5 2 T w o C e l l L i P o l y 5 2 M o t o r B a t t e r y T o t a l T o t a l 2 3 8 T o t a l 1 1 2 6 6 O ne m a j or i s s ue t ha t s t i l l r e m a i ne d w a s t he p r obl e m s w e w e r e ha vi ng w i t h t he M P X 4115A a l t i t ude s e ns or W e w e r e ne ve r a bl e t o r e l i a bl y i nc r e a s e i t s s e ns i t i vi t y, e ve n w he n us i ng t he a m pl i f i c a t i on c i r c ui t s T hi s w a s a p r obl e m t ha t ha d t o be a ddr e s s e d a s w e m ove d on t o de ve l opm e nt of R e v 5. F or one t hi ng w e ne e de d t o de t e r m i ne a w a y t o de t e c t s m a l l c ha nge s i n t he a l t i t ude o f t he M A V on t he or de r of a f e w f e e t t o be a bl e t o de ve l op a r obus t c ont r ol l e r I t a ppe a r e d t ha t w he n pow e r e d on t he s a m e s uppl y a s t he di gi t a l c om pone nt s t he 10 b A D C on t he M e ga 12 8 w a s not a de qua t e f or r e a di ng t he s e ns or out put of t he M P X 4115A T hi s m e a nt t ha t w e m i ght a l s o r un i nt o i s s ue s a s w e i nt e gr a t e d a na l og i ne r t i a l s e ns or s i nt o t he s ys t e m a s w e l l T he r e f or e t he a na l og s ys t e m of

PAGE 56

45 t he M A V 128 w a s a m a j or i s s ue t ha t w oul d ha ve t o be a ddr e s s e d a s w e m ove d i nt o t he de ve l opm e nt of t he R e vi s i on 5 F l i ght S ys t e m

PAGE 57

46 C H A P T E R 4 R E V I S I O N 5: F L I G H T S Y S T E M W I T H O N B O A R D I M U G P S A N D C O N T R O L L E R T he R e vi s i on 4 s ys t e m ha d be e n de bugge d a nd w a s be i ng us e d i n t e s t f l i ght s a nd de ve l opm e nt of a n a ut onom o us c ont r ol l e r ut i l i z i ng t he vi s i on s ys t e m a nd G P S w a s now unde r w a y. W e t he r e f o r e t u r ne d our a t t e nt i on t o de ve l opi ng t he R e vi s i on 5 f l i ght s ys t e m T he ove r a l l a r c hi t e c t ur e of t he s ys t e m r e m a i ne d t h e s a m e w i t h t he A C 4490 500 be i ng us e d f or t he D a t a l i nk a nd t he G H 80D pr ovi di ng n a vi ga t i ona l da t a H ow e ve r t he M A V 128 r e c e i ve d a m a j o r upg r a de w i t h t he ne c e s s a r y c ha nge s m a de t o s uppor t a n I M U s ys t e m onboa r d a s w e l l a s c ont i nue w i t h a l l of i t s o t he r f unc t i ons T he i s s ue s w e f a c e d i n de ve l opi ng t he M A V 128 R 5 a c t ua l l y w e r e r a t he r s i m i l a r t o t hos e f a c e d i n de ve l opi ng t he R e vi s i on 4 s ys t e m A nd a s be f or e i t t ook t hr e e ve r s i ons of t he M A V 128 R 5 be f or e w e go t a l l of t he m a j or pr o bl e m s s or t e d out W e ha d be e n a s ke d t o f ur t he r de c r e a s e t he w e i ght o f t he ove r a l l f l i ght s ys t e m T he onl y s ol ut i on w a s t o no l onge r us e a s e pa r a t e ba t t e r y t o pow e r t he e l e c t r oni c s but i ns t e a d r un of f of t he s a m e ba t t e r y t ha t pow e r e d t he M A V m ot or s T he m a j or i s s ue w i t h t hi s w a s t ha t t he i nput vol t a ge t o a l l r e gul a t or s ha ndl i ng t he ba t t e r y s uppl y w a s now 12V i ns t e a d of 7. 4 T hi s r e s ul t e d i n m uc h hi ghe r t he r m a l di s s i pa t i on r e qui r e m e nt s t he n ha d be e n e xpe r i e nc e d w i t h t he t w o c e l l L i P ol y ba t t e r i e s W e a l s o w a nt e d t o e ns ur e t ha t t he M A V 128 R 5 r e m a i ne d t he s a m e s i z e a s t he R 4C H ow e ve r w e w e r e a ddi ng a l a r ge num be r of ne w c om pone nt s t o s uppor t t he I M U T he r e f or e s om e c ha n ge s t o t he ove r a l l l a yout w e r e m a de W he r e a s pr e vi ous f l i ght c om put e r s us e d t he A T M E G A 128 16 A I w hi c h c a m e i n a T hi n Q ua d F l a t P a c k ( T Q F P )

PAGE 58

47 pa c ka ge w e s t a r t e d us i ng t he A T M E G A 128 16 M I w i t h t he R e vi s i on 5 c om put e r s T hi s de vi c e c a m e i n a M i c r o L e a d F r a m e ( M L F ) pa c ka ge 1/ 3 t he s i z e of t he Q F P T o f ur t he r c ons e r ve boa r d s pa c e t he m i c r o he a de r s pr ovi di ng a c c e s s t o a l l of t he i ndi vi dua l por t s of t he M E G A 128 w e r e r e m ove d W e f e l t t ha t t he y no l onge r ha d a ny r e a l pur pos e due t o t he f a c t t ha t w e ha d not r e a l l y us e d t he m a f t e r a ba ndoni ng t he da ught e r boa r d c onc e pt a f t e r t he M A V 128 R 3. S i nc e t he n r e de s i gns of t he boa r ds ha d be e n us e d t o i nc or por a t e ne w f unc t i ona l i t y i nt o t he M A V 128. F ur t he r m or e t he r e w a s a l s o t he f a c t t ha t m a ny o f t he por t f unc t i ons w e r e a l r e a dy be i ng ut i l i z e d by t h e M A V 128 ha r dw a r e T he num be r of a va i l a bl e pe r i phe r a l s i ns i de t he pr oc e s s or w a s r a pi dl y de c r e a s i ng, a nd s o t he r e w a s ne ve r a ny r e a l r e a s on t o c onne c t a de vi c e t o m os t of t he m i c r ohe a de r s W e a l s o e ve nt ua l l y be ga n t o us e t he J S T Z H / Z R c onne c t or s ys t e m f o r t he s e r vo c a bl e s U s i ng t he s e por t s on t he M A V 128 s a ve d a l ot of r oom due t he f a c t t ha t t he y w e r e ha l f t he s i z e of t he ol d c onne c t or s t hough t he dr a w ba c k w a s t ha t a ny M A V t ha t w a s t o be f l ow n by t he M A V 128 ha d t o ha ve i t s s e r vos m odi f i e d t o us e t he J S T c onne c t or s i ns t e a d of t he s t a nda r d 0. 1 pi t c h he a de r s E ve n a f t e r w e ha ndl e d a l l of t he pr obl e m s w i t h t he s e t r a ns i t i ons t he r e w a s s t i l l t he m a t t e r of de ve l opi ng a n a na l og pr oc e s s i ng s ys t e m f or t he i ne r t i a l s e ns or s t ha t w a s a c c ur a t e e nou gh f or t he c ont r ol l e r D ur i ng t he de v e l opm e nt of t he R e vi s i on 4 f l i ght c om put e r w e ha d i m m e ns e di f f i c ul t y i n c r e a t i ng a pr oc e s s i ng s ys t e m j us t f or t he M P X 4115A a l t i t ude s e ns or a nd w e r e not a bl e t o o bt a i n t he s e ns i t i vi t y t ha t w a s ne e de d. N ow i ns t e a d of j us t one s e ns or t o de a l w i t h, w e ha d t o de a l w i t h t e n a c c e l e r a t i on i n a l l t hr e e di r e c t i ons ( x y z ) a ngul a r r a t e s a r ound a l l t hr e e a xe s ( r ol l pi t c h a nd ya w ) a nd pr e s s ur e s e ns or s t o obt a i n bot h a l t i t ude a nd a i r s pe e d.

PAGE 59

48 T h e M A V 128 R 5 P ow e r S ys t e m T he i ni t i a l goa l of t he pow e r s ys t e m f or t he M A V 1 28 R 5 w a s t o s e pa r a t e t he a na l og a nd di gi t a l pow e r s ys t e m s i n t he hope s of i m pr ovi ng t he pe r f o r m a nc e of t he a na l og pr oc e s s i ng s ys t e m T he r e f or e t he M A V 128R 5A s how n i n F i gu r e 4 1 us e d s e pa r a t e vol t a ge r e gul a t or s f o r t he di gi t a l a nd a na l og s ys t e m s T he T I R E G 113N A r e gul a t o r us e d i n t he M A V 128 R 4C f or G P S pow e r w a s ut i l i z e d o nc e a ga i n, s i nc e i t w a s a va i l a bl e i n 5V ve r s i ons a s w e l l O ne r e gul a t or pow e r e d a l l of t he 5V di gi t a l s ys t e m s a nd t he a l t i m e t e r w hi l e bot h t he g y r os c ope s a nd a c c e l e r om e t e r s e a c h ha d t he i r ow n 5V R E G 113 t o pr ovi de pow e r T hi s w a s a pr e c a ut i on t o t r y a nd ke e p t he s e ns or s i s ol a t e d f r om one a not he r A R E G 113 3. 3V r e gul a t or onc e a ga i n pr ovi de d pow e r f o r t he G P S H ow e ve r a l l of t he r e gul a t or s r e c e i ve d t he i r i nput vo l t a ge di r e c t f r om t he ba t t e r y. T hough t he R E G 113 w a s a s m a l l de vi c e a nd not c a pa bl e of di s s i pa t i ng m or e t he n a w a t t of t he r m a l r e qui r e m e nt s i t c oul d s t i l l ha ndl e t he 7. 4V of t he t w o c e l l ba t t e r y. T he e xc e pt i on w a s t he A C 4490. D ue t o t he f a c t t h a t i t r e qui r e d a l m os t 0 5A t he r e w a s no w a y a R E G 113 c oul d pr ovi de i t pow e r F ur t he r m or e t he L M 3940 us e d pr e vi ous l y c oul d not ha ndl e 7. 4V a s a n i nput a s i t w a s t a r ge t e d onl y f o r c onve r t i ng 5V t o 3. 3V A ne w vol t a ge r e gul a t or ha d t o be f ound. T he s ol ut i on w a s t he L M S 8117A A l s o a va i l a bl e i n a S O T 223 pa c ka ge t he de vi c e w a s ba r e l y c a pa bl e of di s s i pa t i ng t he he a t w e ne e de d. H ow e ve r w e i nc r e a s e d i t s a bi l i t y t o t r a ns f e r he a t t o t he r e s t o f t he boa r d a nd F i gur e 4 1: M A V 128 R 5A

PAGE 60

49 t he s ur r oundi ng a i r by i n c r e a s i ng t he s i z e of t he pa d t ha t i t s m a i n gr ound t a b i s c onne c t e d t o, a nd t he n r e m ovi ng t he pr ot e c t i ve c ove r i ng ove r t ha t pa d f r o m t he P C B de s i gn ( F i gur e 4 2) W e t he n s e nt out t he M A V 128 R 5A de s i gn t o be f a br i c a t e d. T he f i r s t boa r ds c a m e ba c k a nd w e r e a s s e m bl e d r i ght be f o r e w e w e r e t ol d t ha t w e ha d t o c ut t he w e i ght of t he f l i ght s ys t e m e ve n m or e i n R 5. H ow e ve r t hi s w a s e xt r e m e l y di f f i c ul t A f e w gr a m s m i ght be t r i m m e d f r om t he M A V 128, or w e m i ght pos s i bl y be a bl e t o f i nd a s m a l l e r G P S uni t ( t hough us i ng one m i ght r e s ul t i n de c r e a s e d pe r f or m a nc e ) but ne i t he r a t t e m pt w oul d r e a l l y m a ke a di f f e r e nc e i n t he ove r a l l w e i ght of t he f l i ght e l e c t r oni c s T he 26 gr a m a nt e nna us e d f or da t a t r a ns m i s s i on w a s a ge ne r i c de vi c e obt a i ne d f r om A e r oc om m a nd ha d a r ubbe r s hr oud c ove r i ng t ha t w a s not ne c e s s a r i l y ne e de d f o r our a ppl i c a t i on. W e m i ght ha ve be e n a bl e t o r e duc e t he w e i ght by ge t t i ng a c us t om a nt e nna de s i gne d, but e ve n t he n w e w oul d onl y be s a vi ng a bout 10 gr a m s T he r e f or e our s ol ut i on w a s t o r e m ove t he de di c a t e d e l e c t r oni c s ba t t e r y f r om t he e qua t i on, w hi c h t ook up a l m os t ha l f o f t he w e i ght of t he f l i ght s ys t e m H ow e ve r a s a r e s ul t t he e l e c t r oni c s ha d t o be pow e r e d o f f of t he t hr e e c e l l m ot or ba t t e r y. T he pow e r s ys t e m ha d ne ve r be e n de s i gne d t o be c a pa bl e of h a ndl i ng a ba t t e r y i nput o f 12V A s a t e s t w e hooke d up a 12V ba t t e r y t o t he M A V 128 R 5A boa r d, w hi c h ha d a l r e a dy be e n ve r i f i e d t o w or k a t 7. 4V W i t hi n a m i nu t e a l l o f t he vol t a ge r e gul a t or s c onne c t e d t o t he F i gur e 4 2: M A V 128 R 5A A C 4490 R e gul a t or

PAGE 61

50 ba t t e r y w e nt i nt o t he r m a l s hut dow n. W e t he r e f or e ha d t o c om pl e t e l y r e de s i gn t he pow e r s ys t e m on t he M A V 128 R 5A t o s uppor t ope r a t i on of f of t he m ot or ba t t e r y. W e de c i de d t o us e t he s a m e r e gul a t or f o r bot h t he a c c e l e r om e t e r s a nd t he gyr os c ope s a s i t di dn t a ppe a r t h a t a ny r e a l pe r f or m a nc e ga i n w a s be i ng pr oduc e d f r om s e pa r a t i ng t he t w o s e ns or s T he r e f or e w e ne e de d t w o 5V r e gul a t or s one f or a na l og de vi c e s a nd one f or di gi t a l T he 3. 3V r e gul a t or t ha t w a s t o pr ovi de pow e r t o t he G P S c oul d e xi s t unde r ne a t h t he di g i t a l 5V r e gul a t or e s s e nt i a l l y a c t i ng a s a s e c ond s t a ge de vi c e W e s t i l l ha d t he i s s ue of t he 3. 3V r e gul a t or f or t he A C 4490. I f i t w e r e a c ha l l e nge t o e na bl e a 5V r e gul a t or t o r e c e i ve 12V a nd s our c e a l a r ge a m ount of c ur r e nt t he n i t w oul d be e ve n m or e di f f i c u l t t o do s o w i t h a 3 3V r e gul a t or T he r e f o r e i t w a s de c i de d t ha t t he vol t a ge r e gul a t or f or t he A C 4490 s houl d n ot be c onne c t e d t o t he ba t t e r y di r e c t l y, but i ns t e a d r e c e i ve i t s pow e r f r om a not he r vo l t a ge r e gul a t or e s s e nt i a l l y a c t i ng a s a s e c ond s t a ge de vi c e E ve n w i t h t hi s l a yout w e s t i l l r e qui r e d s e ve r a l vol t a ge r e gul a t or s c a pa bl e of di s s i pa t i ng a l ot of t he r m a l e ne r gy, i nc l udi ng bot h t he a na l og 5V r e gul a t o r s i nc e i t w oul d ha ve t he ba t t e r y f or a n i nput a nd t he A C 4490 r e gu l a t or s i nc e i t w oul d ha ve t o s our c e a l a r ge a m ount of c ur r e nt T he di gi t a l 5V r e gul a t or w oul d ha ve t o ha ndl e bot h of t he s e s i t ua t i ons s i nc e i t w oul d be pr ovi di ng t he A C 4490 vol t a ge r e gul a t or w i t h pow e r H ow e ve r t he onl y w a y t ha t w e c oul d e ns ur e t ha t t he s e de vi c e s c oul d ope r a t e a nd not g o i nt o t he r m a l s hut dow n w a s t o m ove t o l a r ge r pa c k a ge s t he n t he S O T 223 c ur r e nt l y be i ng us e d. U s i ng t hr e e s uc h vol t a ge r e gul a t or s w oul d t a ke up a l ot of r e a l e s t a t e on t he M A V 128 boa r d.

PAGE 62

51 T he s ol ut i on w a s t o us e one s i ngl e f i r s t s t a ge vol t a ge r e gul a t or t o s uppl y a l l c ur r e nt f or t he ot he r de vi c e s on t he M A V 128 R 5B I t t ook i n t he 12V s our c e f r om t he ba t t e r y a nd pr oduc e d 5. 5V T he R E G 113 r e gul a t or w hi c h ha s a dr opout vol t a ge of l e s s t he n 400 m A c a n e a s i l y pr oduc e 5V or 3. 3V unde r t he s e c o ndi t i ons B y m ovi ng t o t he R E G 113E A w hi c h us e d a M S O P pa c ka ge t he r e w a s a n i nc r e a s e i n s i z e of 167% bu t a s i gni f i c a nt i nc r e a s e i n t he r m a l di s s i pa t i on, a s c a n b e s e e n i n F i gur e 4 4. C onne c t i ng a l l of t he gr ound pi ns o f t he de vi c e t o a s m a l l gr ound pl a ne t ha t w a s t he n e xpos e d t o t he a i r w a s f ur t he r i ns ur a nc e a ga i ns t t he de vi c e s ove r he a t i ng. U nde r t hi s s e t up, t he R E G 113 c oul d pow e r t he G P S di g i t a l a nd a na l og s ys t e m s T he A C 4490 V ol t a ge R e gul a t or how e ve r r e qui r e d f a r t oo m uc h c ur r e nt t o us e t he R E G 113 V ol t a ge R e gul a t or or e ve n t he L M 3940 or L M S 8117A us e d i n pr e vi ous s ys t e m s N ow t ha t t he i nput t o t he r e gul a t or w a s 5. 5V e ve n t he l a t t e r de vi c e w a s i nc a pa bl e of di s s i pa t i ng t he t he r m a l e ne r gy r e s ul t i ng f r om pr ovi di ng pow e r t o t he R F 2 L M 1 0 8 5 D a t a s h e e t N a t i o n a l S e m i c o n d u c t o r 2 0 0 3 3 R E G 1 1 3 D a t a s h e e t T e x a s I n s t r u m e n t s 2 0 0 3 F i gur e 4 3: T O 263 M a xi m um P ow e r D i s s i pa t i on 2 F i gur e 4 4: S O T 23 M a xi m um P ow e r D i s s i a pt i on 3

PAGE 63

52 T r a ns c e i ve r T he r e f or e t he r e w e r e t w o vol t a ge r e gul a t or s t ha t ne e de d t o be upgr a de d t o a l a r ge r pa c ka ge U s i ng t he a na l ys i s i n T a bl e 4 1, w e de t e r m i ne d t ha t t he s ol ut i on w a s t o us e de vi c e s f r om N a t i ona l s L M 1085 a nd L M 1086 s e r i e s of r e gul a t o r s T he s e de vi c e s w hi c h m e t vol t a ge a nd c ur r e nt r e qui r e m e nt s w e r e a va i l a bl e i n t he T O 263 s ur f a c e m ount pa c ka ge T he pa c ka ge i s ov e r t hr e e t i m e s t he s i z e of t he S O T 223 us e d pr e vi ous l y, but i t i s a l s o a bl e t o di s s i pa t e t w i c e a s m uc h pow e r a s w e l l ( F i gur e 4 3 F i gur e 3 5 ) T he r e f o r e a L M 1086 3 3V r e gul a t or w hi c h c oul d s our c e up t o 1. 5A w a s us e d t o pow e r t he A C 4490 T he L M 1085 c oul d s our c e up t o 3A a nd w a s t he r e f or e w e l l s ui t e d t o pow e r t he r e s t of t he M A V 128 S i nc e t he f i r s t s t a ge r e gul a t or r e qui r e d a non s t a nda r d out put of 5. 5V a n a dj us t a bl e ve r s i on of t he L M 1085 w a s us e d. T he e vol ut i on o f t he pow e r s ys t e m ove r t he de ve l o pm e nt of t he R e vi s i on 5 M A V 128 c a n be s e e n i n F i gur e 4 7. W i t h t he s e c ha nge s t he R e vi s i on 5B w a s s uc c e s s f ul a t pow e r i ng t he s ys t e m w he n c onne c t e d t o a t h r e e c e l l ba t t e r y ( F i gur e 4 5) T he f i r s t s t a ge a nd R F t r a ns c e i ve r r e gul a t or s c oul d di s s i pa t e t he h e a t ge ne r a t e d, t hough t he t e m pe r a t ur e of t he de vi c e s i nc r e a s e d t o ne a r 100C T he de vi c e s t he m s e l ve s c oul d ha ndl e t hi s he a t but t he y i nc r e a s e d t he t e m pe r a t u r e of t he a i r a r oun d t he m s i gni f i c a nt l y T he he a t w a s t oo m uc h f or t he A C 4490, w hi c h w a s c onne c t e d t o t he M A V 128 on t he bot t om of t he boa r d F i gur e 4 5: M A V 128 R 5B

PAGE 64

53 a nd t he r e f or e r i ght ne x t t o t he L M 1086I S 3. 3 L ong t e r m ope r a t i on r e s ul t e d i n da m a ge t o t he R F t r a ns c e i ve r w hi c h no l onge r c om m uni c a t e d a nd ha d t o be r e pl a c e d. B e c a us e of t hi s i s s ue t he l a yout of t he pow e r s ys t e m w a s c ha nge d f or t he M A V 128 R 5C T he c onne c t or f or t he A C 4490 w a s m ove d t o t he oppos i t e s i de of t he boa r d, t o ke e p t h e R F t r a ns c e i ve r a w a y f r om t he he a t of t he vol t a ge r e gul a t or s S i gni f i c a nt s ur f a c e a r e a o f t he M A V 128 w a s a l s o us e d t o c r e a t e e xpos e d c oppe r pl a ne s c onne c t e d t o t he m a i n t a b o f t he T O 263 r e gul a t or s ( F i gu r e 4 6 ) T hi s i nc r e a s e d t he a bi l i t y of t he de vi c e s t o t r a ns f e r he a t t o t he boa r d a nd a m bi e nt a i r a nd t he r e f or e l ow e r e d t he i r i nt e r na l t e m pe r a t u r e T hi s l a yout of t h e pow e r s ys t e m w a s ve r i f i e d on t he be nc ht op t o s uc c e s s f ul l y ope r a t e f o r l ong pe r i ods of t i m e w i t h out s hut t i ng dow n due t o he a t T hough t he r e i s a de t e c t a bl e i nc r e a s e i n t e m pe r a t ur e a r ound t he boa r d i t ha s ne ve r r i s e n t o t he po i nt w he r e t he A C 4490 i s a f f e c t e d. T he M A V 128 R 5C ( F i gur e 4 8) boa r d ha s s i nc e be e n s uc c e s s f ul l y f l ow n on a 24 M A V a nd t he pow e r s ys t e m w a s a bl e t o ope r a t e unde r f l i gh t c ondi t i ons F i gur e 4 6: M A V 128 R 5C P ow e r S ys t e m : F i r s t S t a ge a nd A C 4490 R e gul a t or s

PAGE 65

54 T a bl e 4 1: A na l ys i s of T he r m a l D i s s i pa t i on I s s ue s i n P ow e r i ng t he M A V 128R 5, G H 80 a nd A C 4490 V ol t a ge R e gul a t or T he r m a l D i s s i pa t i on P T = I ( V O U T V IN ) A C 4490 500 ( 3. 3V ) C ur r e nt C ons um pt i on 492 m A M A V 128R 5 ( 5V ) C ur r e nt C ons um pt i on ( D i gi t a l ) 10 m A M A V 128R 5 ( 5V ) C ur r e nt C ons um pt i on ( A na l og) 50 m A G H 80 ( 3 3V ) C ur r e nt C ons um pt i on 88 m A T hr e e C e l l B a t t e r y, M A V 128R 5B / C A C 4490 500, G H 80 F i r s t S t a ge L M 1085 5 5V ( 12V 5 5V ) ( 492 m A + 10 m A + 50 m A + 88 m A = 4. 16W R 5 D i gi t a l R E G 113E A 5 0V ( 5. 5V 5. 0V ) 10 m A = 005W R 5 A na l og R E G 113 E A 5. 0V ( 5. 5V 5. 0V ) 10 m A = 025W A C 4490 L M 3940 3 3V ( 5. 5V 3. 3V ) 492 m A = 1. 0824W G H 80 R E G 113E A 3 3V ( 5. 5V 3. 3V ) 88 m A = 1936W F i gur e 4 7: M A V 128 P ow e r S ys t e m s

PAGE 66

55 D e ve l op m e n t of t h e I n e r t i al an d A n al og C on ve r s i on S ys t e m s W hi l e w e w e r e de t e r m i ni ng t he be s t w a y t o e ns ur e t ha t t he M A V 128 R 5 c oul d s ur vi ve on a t h r e e c e l l ba t t e r y w e w e r e a l s o f oc us i ng on de ve l opi ng t he i ne r t i a l m e a s ur e m e nt s ys t e m t ha t ne e de d t o be i nt e gr a t e d i nt o t he f l i ght c om put e r T he m a j o r c om pone nt s t ha t w e r e r e qui r e d w e r e a c c e l e r om e t e r s t o m e a s ur e t he i ns t a nt a ne ous a c c e l e r a t i on of t he a i r c r a f t i n a l l t hr e e di r e c t i ons a nd gyr os c ope s t o de t e r m i ne t he a ngul a r r a t e s of m ove m e nt a bout a l l t hr e e a xe s T he M P X 4 115A a l s o ne e de d t o be m i g r a t e d t o t he ne w s ys t e m A r e que s t w a s a l s o gi ve n t o i nc l ud e i n R 5 a s e ns or t o m e a s ur e t he a i r s pe e d of t he M A V A f t e r s om e r e s e a r c h, i t w a s de t e r m i ne d t ha t t he M ot or ol a M P X V 4006 pr e s s ur e s e ns or w a s be i ng us e d f or t hi s pur pos e i n ot he r a ut opi l ot s [ 23 ] W e s t i l l ne e de d t o de t e r m i ne a w a y t o a c c ur a t e l y c onv e r t t he a na l og s i gna l s o f a l l o f t he s e de vi c e s i nt o da t a a c c ur a t e e nough t o be us e d i n ou r c ont r ol l e r S i nc e a s a l w a ys s i z e w a s a n i s s ue w e de c i de d t o f oc us on ne w a c c e l e r om e t e r s a nd gyr os c ope s de ve l ope d us i ng M E M S t e c hnol ogy. M a ny I M U s t ha t w e r e i nt e nde d f o r s m a l l pl a t f or m s us e d a c c e l e r om e t e r s f r o m A na l og D e vi c e s i nc l udi ng t he 3D M G W e i ni t i a l l y f oc us e d on us i ng t he A D X L 210 w hi c h ha d a r a nge o f 10G s T he A D X L 210 l i ke m os t of t he A na l og D e vi c e s a c c e l e r om e t e r s w a s dua l a xi s a nd t he r e f o r e w e onl y ne e de d t w o s e ns or s t o m e a s u r e a c c e l e r a t i on i n a l l t hr e e di r e c t i ons T he A D X L 210 pr ovi de d di gi t a l pul s e s f or out pu t s w i t h t he dut y c yc l e de t e r m i ni ng t he a c c e l e r a t i on i t w a s m e a s ur i ng. H ow e ve r t he M e ga 128 onl y ha d t w o i nput c a pt ur e s i gna l s a nd s o t he r e f or e w e w e r e f or c e d t o us e a d i f f e r e nt m e t hod f or r e c or di ng t he da t a F or t una t e l y t he A D X L 210 c oul d a l s o pr ov i de a n a na l og s i gna l w i t h t he a ddi t i on of a c a pa c i t or a nd a n op a m p buf f e r

PAGE 67

56 T he r e w e r e a f e w opt i ons f or t he gyr os c ope s T he 3D M G us e d t he E N C 03J gyr os c ope s f r om M ur a t a H ow e ve r w e w e r e i nf o r m e d t ha t t he y w e r e not a va i l a bl e f or m i l i t a r y a ppl i c a t i ons T oki n a l s o ha d S M D gy r os c ope s a va i l a bl e but t he y w e r e not M E M S ba s e d. I n t he e nd, w e a ga i n w e nt w i t h A na l og D e vi c e s us i ng t he i r A D X R S 300 gyr os c ope T he de vi c e m e a s ur e s a ngul a r r a t e s a t u p t o 300/ s a bout one a xi s O ne m a j or i s s ue w a s t he f a c t t ha t t he de vi c e w a s onl y a va i l a bl e i n a s m a l l ba l l gr i d a r r a y ( B G A ) I t w a s ve r y di f f i c ul t t o a t t a c h t hi s de vi c e t o t he boa r d us i ng t he e qui pm e nt w e ha d a va i l a bl e W e ha d s om e s uc c e s s i n us i ng a S M D S o l de r i ng S t a t i on, f oc us i ng i t s hot ha i r on t he unde r s i de of t he B G A a s w e l ow e r e d i t ont o t he bo a r d, a nd a l l ow i ng t he s ol de r on t he bot t om of t he de vi c e t o f l ow ont o t he pa ds H ow e v e r t hi s p r oc e s s w a s ne ve r c om pl e t e l y r e l i a bl e U s i ng t he s e de vi c e s m e a nt w e ne e de d t w o a c c e l e r om e t e r s a nd t hr e e gy r os c ope s t o c ove r a l l t hr e e a xe s O ne a c c e l e r om e t e r w a s pl a c e d di r e c t l y on t he m a i n boa r d, a nd pr ovi de a c c e l e r a t i on i n t he x a nd y di r e c t i ons A g yr os c ope a l s o w a s pl a c e d on t he boa r d t o pr ovi de t he ya w r a t e T he r e w e r e a l s o t w o da ug ht e r boa r ds w i t h one c ont a i ni ng a n a c c e l e r om e t e r a nd gyr os c ope t o p r ove a c c e l e r a t i on a l ong t he z a xi s a s w e l l a s t he ya w r a t e T he ot he r da ught e r boa r d c ont a i ne d j us t a gyr os c ope t o pr ovi de t he pi t c h r a t e W e ha d de s i gne d t he M A V 128 R 5A w i t h t he goa l of ke e pi ng t he a c c e l e r om e t e r a nd gyr os c ope s e ns or s i s ol a t e d f r om t he di gi t a l pow e r s ys t e m A t t he t i m e w e w e r e a l s o s t i l l w or ki ng on t he M A V 128 R 4C a t t e m pt i ng t o de ve l op a c i r c ui t boa r d us i ng op a m ps t o i nc r e a s e t he s e ns i t i vi t y of t he s e ns or T he r e f o r e l i k e t he R e vi s i on 4C t he ne w s ys t e m i nc l ude d a c onne c t or t o a s e pa r a t e c i r c ui t f or a m pl i f yi ng t he a l t i t ude da t a

PAGE 68

57 B y t he t i m e w e s t a r t e d de ve l opi ng t he M A V 128 R 5B t o c or r e c t t he i s s ue s i n t he pow e r s ys t e m w e ha d a l s o r e a l i z e d t hr ough our e xpe r i m e nt a t i on w i t h t he a l t i t ude s e ns or t ha t pa r t of t he pr obl e m w a s t he a f f e c t t he M e ga 128 pr oc e s s or c or e ha d on t he i nt e r na l A D C T he r e f o r e w he n w e r e de s i gne d t he pow e r s ys t e m f or t he M A V 128 R 5B w e a l s o m a de a n a t t e m p t t o f ur t he r i s ol a t e t he a na l og s ys t e m s f r om t he di gi t a l T he A T m e ga 128 d i d pr ovi de s e pa r a t e pow e r a nd gr ound c onne c t i ons f or t he i nt e r na l A D C T he s e s i gna l s w e r e t i e d t o a na l og pow e r a nd gr ound a s w e r e a l l of t he i ne r t i a l s e ns or s A s di s c us s e d e a r l i e r t he di gi t a l a nd a na l og pow e r s ys t e m s us e d di f f e r e nt vol t a ge r e gul a t or s A nd s t a r t i ng w i t h t he R 5B s ys t e m w e ke pt t he a na l og a nd di gi t a l gr ounds s e pa r a t e f r om one a not he r A s s oon a s gr ound e nt e r e d i n t o t he boa r d a t t he m a i n pow e r c o nne c t or i t c onne c t e d di r e c t l y t o t he di gi t a l gr ound I t a l s o c onne c t e d t hr ough a 0 R e s i s t or t o A na l og G r ound. T hi s w a s t he onl y c onne c t i on be t w e e n t he t w o s i gna l s T hi s w a y f l uc t ua t i ons on t he d i gi t a l g r ound pl a ne due t o t he pr oc e s s or ope r a t i on w e r e l e s s l i ke l y t o a f f e c t t he a na l og s ys t e m s A not he r c ha nge m a de dur i ng t he de ve l opm e nt of t he M A V 128 R 5B w a s a m odi f i c a t i on of t he l a yout of t he i ne r t i a l s e ns or s I n t he pr e vi ous de s i gns one da ught e r boa r d i nc l ude d a n a c c e l e r om e t e r a nd gyr os c ope a nd ot he r boa r d j us t ha d a gyr o T o r e duc e t he c om pl e xi t y o f t he s ys t e m a nd m a ke m a nuf a c t ur i ng e a s i e r a s i ngl e da ught e r F i gur e 4 8: M A V 128 R 5C

PAGE 69

58 boa r d w a s de s i gne d c ont a i ni ng bot h a n a c c e l e r om e t e r a nd a gy r os c ope T w o s uc h da ught e r boa r ds w e r e us e d w i t h t he s ys t e m a nd pr ovi de a l l of t he i ne r t i a l da t a e xc e pt t he ya w r a t e T hi s w a s a ga i n c ove r e d by one gyr os c op e t ha t r e m a i ne d on t he m a i n boa r d W e a l s o w e nt f r om us i ng t he A D X L 210 a c c e l e r om e t e r s t o t he m o r e a c c ur a t e A D X L 203s a ne w s e ns or f r om A na l og D e vi c e s I t a l s o pr ovi de d a na l og out put s i ns t e a d of P W M s i gna l s m e a n i ng t he op a m p buf f e r s w e r e no l onge r ne e de d. H ow e ve r w hi l e t he A D X L 210 ha d a r a nge of 10G s t he 203 onl y ha d a r a nge o f 1. 7G s S om e o f t he i s s ue s w i t h t he 3D M G i n t he e a r l i e r t e s t s ha d be e n t he s a t ur a t i on of i t s a c c e l e r om e t e r s unde r c e r t a i n f l i ght c on di t i ons a nd t hos e de vi c e s ha d a s i m i l a r r a nge t o t he A D X L 203. H ow e ve r t he s e ns or w a s now be i ng us e d on t he P oc ke t M A V C ont r ol l e r w i t h s om e s uc c e s s T he r e f or e w e de c i de d t o us e t he s e s e ns or s f or t he M A V 128 R 5B a s w e l l O ne be ne f i t of t he ne w l a yout of t he i ne r t i a l s e ns or s w a s t ha t i f t he r e w e r e p r obl e m s w i t h t he A D X L 203, i t w oul d be e a s y t o go ba c k t o t he ol d a c c e l e r om e t e r s by j us t r e pl a c i ng t he da ught e r boa r ds W i t h t he a na l og a nd di gi t a l s ys t e m s c om pl e t e l y i s o l a t e d, w e w e r e hope f ul t ha t t he a m pl i f i e r s ys t e m de ve l ope d pr e vi ous l y w oul d be e nough t o obt a i n a s e ns i t i vi t y of a f e w f e e t f r om t he M P X 4115A T he r e f or e t he a m pl i f i c a t i on c i r c ui t w a s m i g r a t e d di r e c t l y i nt o t he M A V 128 de s i gn. H ow e ve r t e s t s of t he M A V 1 28 R 5B s how e d no f ur t he r i m pr ove m e nt i n t he r e s ol ut i on of t he a l t i t ude s e ns or W e w e r e ge t t i ng da t a f r o m t he i ne r t i a l s e ns or s but t he r e w a s no r e a l w a y t o t e l l i f t he r e s ul t s w e r e a c c ur a t e e nough f o r a ut onom ous c ont r ol I t w a s t he r e f or e pos s i bl e t ha t t he s a m e pr obl e m s w e w e r e f a c i ng w i t h t he a l t i t ude s e ns or w e r e oc c ur r i ng w i t h t he ot he r s e ns or s a s w e l l

PAGE 70

59 B e c a us e of t he s e i s s ue s w e de c i de d t o s e a r c h f o r a hi gh r e s ol ut i on e xt e r na l A D C t ha t w oul d i nt e r f a c e t o t he M e ga 128. W e hope d t o f ur t he r i s ol a t e t he a na l og s ys t e m s f r om t he M e ga 128, a nd a c hi e ve gr e a t e r s e ns i t i vi t y i n a l l of t he s e ns or da t a T he M A V 128 R 5C w a s de s i gne d t o us e t hi s ne w A D C t o c onve r t t he a na l og s i gna l s of t he i ne r t i a l s e ns or s a s w e l l a s t he a l t i t ude a nd a i r s pe e d pr e s s ur e s e ns or s T he M e ga 128 A D C w a s s t i l l us e d f or l ow r e s ol ut i on de vi c e s T h i s i nc l ude d t he s e r vo f e e dba c k s ys t e m s a s w e l l a s a s i m pl e r e s i s t or di vi de r c i r c ui t t o dr op t he ba t t e r y t o a l e ve l t ha t c oul d be r e a d by t he M e ga 128. T hi s a l l ow e d us t o m oni t or t he ba t t e r y vol t a ge a nd l a nd t he M A V i f t he ba t t e r y w a s l ow F i na l l y, t he M e g a 128 a l s o r e a d a t e m pe r a t u r e s e ns or t ha t w a s i nc l ude d i n t he gyr os c ope S i nc e t he i ne r t i a l a nd pr e s s ur e s e ns or s w e r e a f f e c t e d by t e m pe r a t ur e c ondi t i ons i t w a s de c i de d t o i nc l ude t he s i gna l i n c a s e i t w a s ne e de d f or f ut ur e c ont r ol l e r de ve l opm e nt O ur r e qui r e m e nt s f or a n e xt e r na l A D C i nc l ude d a s e r i a l i nt e r f a c e t hr ough e i t he r t he S P I P o r t o r t he I 2C B us W e di d not w a nt t o de a l w i t h ha vi ng t o c onne c t s e ve r a l t r a c e s be t w e e n t he M e ga 128 a nd t he A D C w hi c h c oul d m a ke t he l a yout o f t he boa r d m o r e di f f i c ul t a nd s o A D C s w i t h pa r a l l e l i nt e r f a c e s w e r e not c ons i de r e d. W e ne e de d a t l e a s t 8 c ha nne l s t o s uppor t t he a c c e l e r om e t e r s gy r os c ope s a nd t he p r e s s ur e s e ns or s t hough t he s e ns or s c oul d pot e nt i a l l y be di vi de d i n t o m ul t i p l e A D C s i f ne c e s s a r y. I t w a s f e l t t ha t us i ng a 14 bi t or l e s s A D C w oul d m os t l i ke l y r e s ul t i n not e nough s e ns i t i vi t y f o r ou r ne e ds ba s e d on t he pr e vi ous e xpe r i m e nt s F ur t he r m or e i t w a s f ound t ha t of t e n c om pa ni e s of f e r t he s a m e de vi c e i n 12 14 o r 16 bi t r e s ol ut i ons w i t h t he onl y di f f e r e nc e be i ng t he c os t of t he de vi c e s T he r e f or e w e on l y l ooke d f or A D C s w i t h a t l e a s t 16 bi t s o f r e s ol ut i on.

PAGE 71

60 I n t he e nd w e w e nt w i t h e ve n h i ghe r r e s ol ut i on i n c hoos i ng t he T I A D S 1256, a n 8 C ha nne l 24 bi t S e r i a l A D C f or t he M A V 128 R 5C W i t h t hi s m uc h a c c ur a c y, a ny i s s ue s i n t he da t a w e r e m os t l i ke l y t o r e s ul t f r om t he s e ns or s t he m s e l ve s or i nt e r f e r e nc e f r om ot he r de vi c e s not t he A D C H ope f u l l y, t he M A V 128 R 5C de s i gn ha d a l r e a dy de a l t w i t h t he s e i s s ue s t hough t he s e l e c t i on of t he s e ns or s a nd t he i s ol a t i on o f t he pow e r s ys t e m s T he A D S 1256 i n t e r f a c e d t o t he M e ga 128 t hr ough t he S P I P or t I t i nc l ude d m ul t i pl e f e a t ur e s i nc l udi ng a l ow pa s s f i l t e r a pr og r a m m a bl e ga i n a m pl i f i e r a nd a l s o a n i nput buf f e r t ha t he l pe d pr oduc e t he l a r ge r e s ol ut i on f or t he A D C H ow e ve r t he i np ut buf f e r c oul d onl y f unc t i on w i t h i nput s of l e s s t he n 3V T h e a c c e l e r om e t e r s a nd gyr os c ope s pr oduc e 2. 5V unde r nul l m e a s ur e m e nt s but c oul d e a s i l y go a bove t hi s l i m i t F ur t he r m or e t he m e a s ur e m e nt of t he a l t i t ude s e ns or on t he gr ound w a s a bout 4. 1V W e t h e r e f or e ha d t o di s a bl e t he i npu t buf f e r t hough w e onl y s a c r i f i c e d one o r t w o bi t s of r e s ol ut i on a s a r e s ul t O nc e t he de t a i l s of i n t e r f a c i ng t he A D S 1256 t o t he i ne r t i a l s e ns or s a nd t he M e ga 128 ha d be e n de t e r m i ne d, t he M A V 128 R 5C de s i gn w a s s e nt t o be f a b r i c a t e d. W he n t he boa r ds w e r e r e t ur ne d, t he pa r t s w e r e s ol de r e d ont o t he s ys t e m H ow e ve r ou r m e t hod of a t t a c hi ng t he gyr o B G A s t o t he P C B s f a i l e d t hi s t i m e T he Y a w G yr os c ope onl y pr oduc e d a s i gna l w he n i t w a s he a t e d up by t he a i r gun of t he S M D s ol de r i ng s t a t i on, a nd t he n on l y f o r a f e w m i nut e s O bvi ous l y, t he r e w a s a n i nt e r m i t t e nt c onne c t i on on t he bot t om of t he de vi c e T he s ol ut i on w a s t o i n s t e a d ha ve t w o c om pl e t e M A V 128 R 5C s ys t e m s a s s e m bl e d by a c ont r a c t or T hi s e ns ur e d t ha t a l l of t he c om pone nt s w e r e a c c ur a t e l y pl a c e d on t he boa r ds

PAGE 72

61 Figure 4-9: MAV128 R5C Software Architecture

PAGE 73

62 W i t h t he M A V 128 R 5C P C B s f a br i c a t e d, a s s e m bl e d, a nd r e t ur ne d t o us w e s e t out t o m odi f y t he c ode s o t ha t not onl y G P S da t a but a l s o i ne r t i a l t e l e m e t r y c oul d be r e c or de d a nd s e nt dow n t o t he g r ound s t a t i on upon r e que s t T he m a j or c ha nge w a s t he a ddi t i on of t he a ds 1256_i nt m odul e w hi c h us e d t he M e ga 128 s S P I P o r t t o c ont r ol t he A D S 1256, a c c e s s t he i ne r t i a l s e ns or da t a a nd s t or e i t t o t he m a v_da t a s t r uc t ur e T he ove r a l l pr ogr a m w a s c ha nge d f r om da t a M A V t o a ut oM A V be c a us e f or t he f i r s t t i m e s i nc e t he e a r l y e xpe r i m e nt s of t he pr ot ot ype f l i ght s ys t e m s w e w e r e goi ng t o a t t e m pt t o ha ve a n onboa r d c ont r ol l e r f l y t he a i r pl a ne di r e c t l y t h r ough t he M A V 128. T he gr ound s t a t i on w a s onl y t o be us e d t o m oni t o r t he s i t ua t i on. O nc e w e ha d f i ni s he d de ve l opi ng t he a ut oM A V pr ogr a m ( F i gu r e 4 9 ) w e s e t out t o t e s t t he M A V 128 R 5C on t he gr ound W e a l r e a dy kne w t ha t t he s e ns or s w e r e s e ndi ng da t a w i t h t he gy r os s how i ng no a ngul a r r a t e s unde r s t a t i c c ondi t i ons F ur t he r m or e t he a c c e l e r om e t e r s c oul d onl y de t e c t t he f or c e o f gr a vi t y. H ow e ve r unt i l f l i ght t e s t i ng be ga n w e di d not know w he t he r t he A D S 1256 a nd i s ol a t e d a na l og pow e r s ys t e m w oul d be e nough t o pr ovi d e a c c ur a t e i ne r t i a l da t a f o r a ut ono m ous c ont r ol W e di d ha ve t he a bi l i t y t o t e s t t he s e ns i t i vi t y of t he a l t i t ude s e ns or t hough. W e c a r r i e d t he M A V 128 R 5C boa r d up a nd dow n s e ve r a l f l i ght s of s t a i r s a nd t he n a na l yz e d t he t e l e m e t r y us i ng a l ow pa s s f i l t e r t o de t e r m i ne t he s e ns i t i vi t y of t he s ys t e m B o t h t he r a w a nd f i l t e r e d da t a a r e s how n i n F i gur e 4 10. W e de t e r m i ne d t ha t t he s e ns or now ha d a s e ns i t i vi t y of a r ound 2. 5 f e e t due t o t he f a c t t ha t w e c oul d a c t ua l l y de t e r m i ne t he di s c r e t e c ha nge s i n a l t i t ude t ha t F i gur e 4 10: M A V 128 R 5C A l t i t ude D a t a

PAGE 74

63 oc c ur r e d f or e ve r y s t e p t a ke n I t a ppe a r e d t ha t our i s s ue s i n c onve r t i ng t he a na l og da t a w e r e s ol ve d. W e now ha d t o t e s t t he s ys t e m unde r a c t ua l f l i ght c ondi t i ons t o de t e r m i ne w he t he r t he M A V 128 R 5C c oul d be us e d a s a n I M U t o a ut onom ous l y f l y a M A V F l i gh t T e s t i n g an d O n b oar d C on t r ol l e r D e ve l op m e n t B y t he t i m e w e w e r e r e a dy t o be gi n f l i ght t e s t i ng o f t he M A V 128 R 5C t he s e c ond ve r s i on of t he A V C A A F pl a ne w a s r e a dy ( F i gur e 4 11) T he r e f o r e t he pl a ne w a s m odi f i e d s o t ha t t he s e r vos c oul d be c ont r ol l e d di r e c t l y by t he M A V 128. T hi s m e t hod of c ont r ol ha d not be e n a t t e m pt e d s i nc e t he i ni t i a l t e s t s of t he p r ot ot ype f l i ght s ys t e m a nd ha d be e n a ba ndone d be c a us e of c ont r ol l a bi l i t y i s s ue s H ow e ve r t hi s f unc t i ona l i t y w a s ne c e s s a r y f or de ve l opi ng a n onboa r d c ont r ol l e r s o our f i r s t f l i ght t e s t s w e r e t o ve r i f y t he ope r a t i on of t he f l y by w i r e s ys t e m W e s oon r a n i nt o i s s ue s how e ve r w i t h t he s e r vos i nt e r m i t t e nt l y r ot a t i ng t he c ont r ol s ur f a c e s t o m a xi m um de f l e c t i on A f t e r s om e a na l ys i s i t w a s de t e r m i ne d t ha t t he s e r vo pul s e w hi c h s houl d oc c ur a t a f r e que nc y of 50 H z w i t h a dut y c yc l e o f 5 % t o 10% w a s oc c a s i ona l l y not s w i t c hi ng c or r e c t l y. T he r e a s on f or t h i s w a s t ha t t he s e r vo c ha nne l s w e r e us i ng t he out put c om pa r e f unc t i ons of t he M e ga 1 28. W he n a T i m e r i n t he M e ga 128 r e a c he d a poi nt e qua l t o t he c om pa r e r e gi s t e r t he s e r vo c ont r ol pi n t oggl e d, a nd t he F i gur e 4 11: A V C A A F 2. 0

PAGE 75

64 pr oc e s s or w oul d i nt e r r upt s o t ha t a n i nt e r r up t s e r vi c e r out i ne ( I S R ) f unc t i on c oul d r un T hi s f unc t i on w oul d t he n upda t e t he c om pa r e r e gi s t e r s o t ha t t he s ys t e m w oul d a c t i va t e a ga i n on t he ne xt t r a ns i t i on o f t he s e r vo c om m a nd s i gna l H ow e ve r w i t h s o m a ny di f f e r e nt i n t e r r upt ba s e d s ys t e m s now be i ng us e d i n t he M e ga 128, oc c a s i ona l l y a n I S R w oul d not ha ve a c ha nc e t o r un be f or e t he ne xt t r a ns i t i on, a nd s o t he s e r vo pul s e w oul d r e m a i n a t i t s c u r r e nt l e ve l c a us i ng t he unc om m a nd e d m ove m e nt s T he s ol ut i on w a s t o us e t he M e ga 128 s P W M C ha nne l s a di f f e r e nt pa r t of t he T i m e r S ys t e m s o t ha t t he p r oc e s s or c oul d ge ne r a t e t he s e r vo pul s e s w i t hout r e l yi ng on c ode T hi s ha d pr e vi ous l y not be e n us e d be c a us e us i ng t he P W M s ys t e m r e s ul t e d i n t he l os s of a n i nput c a pt ur e de vi c e H ow e ve r t he pe r i p he r a l i n que s t i on w a s not be i ng us e d, a nd i t w a s ne c e s s a r y t o m ove t o P W M c ont r ol i f t h e f l y by w i r e s ys t e m w a s t o w or k pr ope r l y. T he r e f o r e t he ne c e s s a r y c ha nge s t o t he s e r vo_c ont r ol m odul e w e r e m a de T he m ove t o P W M c ont r o l e l i m i na t e d t he i s s ue s w i t h t he s ys t e m a nd t he f l y by w i r e s ys t e m ha s s i nc e be e n ve r i f i e d. W i t h t he M A V 128 now f l yi ng t he a i r pl a ne w e be g a n s e n di ng i ne r t i a l da t a t o t he gr ound. W i t h t he a ddi t i o n of a 900 M H z ga i n a nt e nna t o t he g r ound s t a t i on, w e w e r e a bl e t o i m pr ove t he r e c e pt i on o f t he da t a l i nk, t hough s t i l l onl y a t a r a t e of a bout 25 H z T hi s a l l ow e d us t o ha ve s om e vi e w i n t o t he ope r a t i on o f t he c ont r ol l e r w hi c h w a s now be i ng c ode d i ns i de of t he a ut oM A V c ont r ol l oop T he a u t oM A V c ode w a s m odi f i e d t o c onve r t t he r e s ul t s f r om t he A D S 1256 t o f l oa t i ng poi nt nu m be r s r e pr e s e nt i ng t he a c t ua l vol t a ge pr e s e nt a t t he i nput s o f t he A D C I t c oul d t he n be c onve r t e d t o t he a c t ua l uni t s of t he s e ns or i n que s t i on, a s s how n i n T a bl e 4 2

PAGE 76

65 D ue t o t he f a c t t ha t e a c h gy r os c ope va r i e d i n i t s nu l l vol t a ge by 2V a ut oM A V on s t a r t up a l s o a ve r a ge d t he f i r s t 10 i ni t i a l r e a di ngs of e a c h gyr os c ope a ve r a ge t he m t oge t he r a nd us e d t he r e s ul t s t o c a l c ul a t e t he nul l vol t a ge of e a c h de vi c e H ow e ve r t hi s a l s o m e a nt t ha t on pow e r up, t he M A V 128 R 5C ne e de d t o r e m a i n s t a t i ona r y T a bl e 4 2: C onve r s i on F or m ul a s f or I ne r t i a l S e ns or s A D S 1256 A D C 0V 8388607 ( 0x7F F F F F ) 2. 5V 0 5V 8388608 ( 0x800000) F l oa t V ol t a ge R e pr e s e nt a t i on C onve r s i on I ne r t i a l S e ns or s A c c e l e r om e t e r s N ul l ( 0G A c c e l e r a t i on) 2. 5V S e ns i t i vi t y 1 V / G G yr os c ope s N ul l ( 0 / s R ot a t i on) 2. 5V S e ns i t i vi t y 005V / / s A l t i t ude S e ns or N ul l ( 0 F e e t ) 4. 1V C onve r s i on ( V ol t a ge t o kP a ) A i r s pe e d S e ns or C onve r s i on ( V ol t a ge t o kP a ) A s i m pl e c ont r ol l e r w a s de ve l ope d by M uj a hi d A b dul r a hi m a nd t he n por t e d i nt o t he a ut oM A V pr ogr a m A m o r e a dv a nc e d but t e r w or t h l ow pa s s f i l t e r w a s us e d t o f u r t he r f i l t e r t he da t a T he n, t he r e s ul t s f r om t he a c c e l e r om e t e r s w e r e us e d t o de t e r m i ne t he gr a vi t a t i ona l ve c t or a nd f r om t ha t da t a a s t a t e e s t i m a t or c oul d obt a i n t he c ur r e nt pi t c h a nd r ol l o f t he M A V T he c o nt r ol l e r t he n us e d a s i m pl e pr opor t i ona l c ont r ol l e r t o or de r t he s e r vos t o pos i t i ons t ha t w oul d a dj us t t he pi t c h a nd r ol l a ngl e s t o 0, us i ng t he a ngul a r

PAGE 77

66 r a t e s f r om t he gyr os t o c ont r ol t he m ove m e nt O ne i ni t i a l r e s ul t t ha t c a m e ou t of t hi s w or k w a s t ha t w e now kne w f o r c e r t a i n t ha t t he M e ga 128 c oul d r un a n i ne r t i a l ba s e d c ont r ol l e r onboa r d W he r e a s a t m i ni m um a ny c ont r ol a l gor i t h m ne e ds t o r un a t a bout 30 H z t he M e ga 128 w a s r unni ng t he c ont r ol l oop a t o ve r 250 H z S t a t i c t e s t s i n t he l a b s how e d t ha t t he s t a t e e s t i m a t or c oul d a c c ur a t e l y de t e r m i ne t he or i e nt a t i on of t he a i r p l a ne H ow e ve r t he f i r s t f l i gh t t e s t s w e r e uns uc c e s s f ul i n t ha t t he M A V 128 c ont r ol l e r c oul d no t s t a bi l i z e t he M A V T e l e m e t r y on t he gr ound i ndi c a t e d t ha t t he i s s ue w a s t he a c c e l e r om e t e r s w hi c h w e r e s ho w i ng e xt r e m e l y noi s y da t a e ve n w i t h t he B ut t e r w or t h l ow pa s s f i l t e r T he a c c e l e r om e t e r s ha d a ha r dw a r e l ow pa s s f i l t e r bui l t i n, w i t h t he c ut of f f r e que nc y de t e r m i ne d by t he va l ue o f a n e xt e r na l c a pa c i t or a t t he out pu t of t he s e ns or W e ha d i ni t i a l l y us e d a 001 F c a pa c i t or r e s ul t i ng i n a 5 kH z ba ndw i dt h A s w e w e r e s e e i ng a l ot of hi gh f r e que nc y noi s e i n t he a c c e l e r om e t e r s i gna l s t he c a pa c i t or w a s c ha nge d t o a 1 F c a pa c i t or i ns t e a d, gi vi ng us a c u t of f f r e que nc y of 50 H z T he noi s e s t i l l r e m a i ne d, how e ve r T he r e w a s s om e t hough t ha t t h e i s s ue s w i t h t he a c c e l e r om e t e r s w a s t ha t t he y w e r e be i ng a f f e c t e d by E M F ge ne r a t e d b y t he M A V dr i ve m ot or H ow e ve r i t w a s c ons i de r e d m or e l i ke l y t o be c a us e d by t he vi b r a t i ons of t he a i r pl a ne a s i t m ove d t hr ough t he a i r T he M A V w a s t he r e f o r e hooke d b a c k up i nt o t he j i g pr e vi ous l y us e d i n t he 3D M G t e s t s O nc e a ga i n, t he s e t e s t s s how e d t ha t vi br a t i on w a s a ke y f a c t or s i nc e t he a c c e l e r om e t e r r e s ul t s r a pi dl y os c i l l a t e d w he ne ve r t he pl a ne w a s be i ng s ha ke n, w he t he r due t o t he d r i ve m ot o r o r i nduc e d m o t i on ( F i gur e 4 12)

PAGE 78

67 A dj us t m e nt s t o t he c ut of f f r e que nc y of t he B ut t e r w or t h f i l t e r ha d l i t t l e a f f e c t W e t he n c a m e t o t he r e a l i z a t i on t ha t t he f i l t e r i t s e l f w a s not r unni ng pr ope r l y T hi s w a s due t o t he f a c t t ha t t he r e s ul t s of t he M e ga 128 ba s e d f i l t e r di f f e r e d f r o m t ha t of a P C ba s e d f i l t e r w he n t he s a m e r a w da t a w a s s uppl i e d t o bot h. W e e ve nt ua l l y de t e r m i ne d t ha t t he i s s ue w a s w i t h t he f r e que nc y o f t he c ont r ol l e r on t he M e ga 128. T he c ont r ol l oop r a n a s qui c kl y a s pos s i bl e w i t h t he r a t e s l i ght l y va r yi ng a s w e i ns e r t e d di f f e r e nt i ns t r uc t i ons t o t r y a nd de bug t he s ys t e m F ur t he r m or e t he r a t e w a s a l s o a f f e c t e d dur i ng t hos e c yc l e s w he r e a pa c ke t ha d be e n r e que s t e d a nd w a s be i ng s e nt t o t h e gr ound s t a t i on. W e de c i de t o m odi f y t he pr ogr a m s o t ha t t he c ont r ol l oop a l w a ys r a n a t a r a t e of 50 H z T he t i m e r i n t he M e ga 128 t ha t w a s be i ng us e d f or c a pt u r i ng s e r vo c om m a nds f r om t he R C s ys t e m w a s F i gur e 4 12: A c c e l e r om e t e r D a t a

PAGE 79

68 us e d a s a r e f e r e nc e t o s t a l l t he c ont r ol l oop a t t he e nd of a c y c l e unt i l a f ul l 20 m s ha d pa s s e d. O nc e t hi s m odi f i c a t i on w a s m a de t he r e s ul t s of t he M e ga 128 ba s e d B ut t e r w or t h f i l t e r m a t c he d t ha t of t he g r ound. H ow e ve r w e s t i l l c oul d not f i l t e r out t he hi gh f r e que nc y os c i l l a t i ons i n t he da t a c a us e d by t he vi b r a t i on o f t he M A V A S a r e s ul t t he c ont r ol l e r w oul d not s e e t he l ow e r f r e que nc y c ha nge s i n t he gr a vi t a t i ona l ve c t or r e s ul t i ng f r om t he m ove m e nt o f t he p l a ne T he f i l t e r s w e a r e e m pl oyi ng a r e s t i l l una bl e t o pr ovi de us w i t h t he ne c e s s a r y da t a t o a l l ow t he s ys t e m t o t r a c k t he gr a vi t a t i ona l ve c t o r

PAGE 80

69 C H A P T E R 5 C O N C L U S I O N S A N D F U T U R E W O R K A t t hi s poi nt w e a r e s t i l l a t t e m pt i ng t o de ve l op a s i m pl e onboa r d c ont r ol l e r f or t he M A V 128 R 5 t ha t w i l l e na bl e i ne r t i a l ba s e d a ut ono m ous c ont r ol A f e w t e s t f l i ght s ha ve be e n f l ow n on a m uc h l a r ge r R C a i r c r a f t T he r e s ul t i ng a c c e l e r om e t e r t e l e m e t r y s how n i n F i gur e 5 1, s how s none of t he hi gh f r e que nc y n oi s e w e ha ve not i c e d on t he M A V f l i ght s T he r e f or e t he i s s ue w i t h vi br a t i on a ppe a r s t o be di r e c t l y r e l a t e d t o t he s i z e of pl a t f or m t ha t w e m us t c ont r ol A f t e r i nve s t i ga t i on, w e ha ve di s c ove r e d t ha t w e a r e not a l one i n our di f f i c ul t i e s t o i s ol a t e t he vi br a t i on of t he a i r c r a f t O t he r pr oj e c t s a t t e m pt i ng F i gur e 5 1: A c c e l e r om e t e r R e s ul t s 6 F oot R C A i r c r a f t P l a t f o r m

PAGE 81

70 t o de ve l op a ut opi l ot s f or a i r c r a f t ha ve of t e n us e d a m i xt ur e of s pe c i a l m e c ha ni c a l m ount s t o da m pe n t he vi b r a t i ons of t he a vi oni c s s ys t e m a s w e l l a s K a l m a n f i l t e r s t o r e t r i e ve t he ne c e s s a r y da t a f r om t he a c c e l e r om e t e r s [ 24 ] H ow e ve r t he f or m e r i s de pe nde nt on t he de s i gn of t he a i r c r a f t a nd t he l a t t e r m us t be de a l t w i t h f r om a c ont r ol s pe r s pe c t i ve A t t hi s poi nt i t a ppe a r s t ha t t he r e i s not hi ng m or e t ha t c a n be done f r om t he pe r s pe c t i ve of ha r dw a r e de ve l opm e nt t o f ur t he r de ve l op t he M A V 128. G i ve n t he r i ght m ount i ng s ys t e m a nd t he pr ope r c ont r ol m e t hods i t w i l l be pos s i bl e f or t he M A V 128 R 5C t o us e i t s i nt e gr a t e d I M U t o s t a bi l i z e t he M A V i n t he a i r a n d us e t he G P S s ys t e m a l r e a dy de ve l ope d i n t he R e vi s i on 4 pl a t f or m t o na vi ga t e W i t h t he c om bi na t i on of t he M A V 128 R 5C I ne r t i a l a nd G P S S ys t e m s a nd t he vi s i on p r oc e s s i ng c a pa bi l i t i e s of t he gr ound s t a t i on t ha t a r e e ve n now s t a r t i ng t o c om e onl i ne t he A V C A A F a i r c r a f t s houl d be a bl e t o be g i n t o pe r f or m i nc r e a s i ngl y c om pl e x m a ne uve r s i n ope n f i e l ds T hi s w oul d not how e ve r ne c e s s a r i l y t r a ns l a t e t o be i ng a bl e t o do s o i n a n u r ba n e nvi r on m e nt W i t h t he i ne r t i a l ba s e d c ont r ol onboa r d, t he g r ound s t a t i on i s no l onge r ne c e s s a r y f or a i r c r a f t s t a bi l i z a t i on. F u r t he r m or e i f t he c ont r ol a l gor i t hm s f or t he G P S ba s e d na vi ga t i ona l s ys t e m s a r e m ove d i nt o t he M e ga 128, t he gr ound s t a t i on i nt e r f a c e c oul d be s i m pl i f i e d t o j us t pr ovi de s t a t us i nf or m a t i on a bout t he s t a t e of t he a i r c r a f t a nd a l l ow t he G P S w a ypoi nt s us e d f or na vi ga t i on t o be upda t e d H ow e ve r s i nc e t he M A V 128 w i l l ne ve r be a bl e t o pr ovi de t he c om put a t i ona l pow e r f or vi s i on pr oc e s s i ng, t hi s w oul d ha ve t o r e m a i n i n t he gr ound s t a t i on A s s uc h, t he M A V w oul d s t i l l ne e d t o s e nd a vi de o s i gna l t o t he gr oun d t o be a na l yz e d, a nd r e c e i ve s om e c ont r ol l e r c om m a nds

PAGE 82

71 I n t he f l i ght t e s t s of t he M A V s t ha t ha ve be e n c onduc t e d i n a n op e n f i e l d, vi de o noi s e oc c a s i ona l l y oc c ur s but t he s ys t e m i s of t e n a bl e t o c om pe ns a t e H ow e ve r i n a n ur ba n e nvi r onm e nt t he r e m a y be s e ve r a l s e c onds or m or e of d r opout s t ha t c oul d ve r y w e l l be f a t a l f or t he s ur vi va l o f t he a i r pl a ne I t w i l l be ve r y d i f f i c ul t t o e ns ur e a r e l i a bl e l i nk be t w e e n t he gr ound s t a t i on a nd t he M A V onc e t he ve hi c l e i s f l yi ng a m ongs t bui l di ngs T he r e f or e t he ne xt s t e p i n t he de ve l opm e nt of t he a vi oni c s f l i ght s ys t e m m us t be t o s t a r t t he pr oc e s s of m i g r a t i ng t he vi s i on pr oc e s s i ng f un c t i ons of t he g r ound s t a t i on i nt o t he M A V T he f unc t i ona l i t y of t he M A V 128 i s of vi t a l i m por t a nc e t o t he a bi l i t y of a M A V t o s ur vi ve i n a n ur ba n e nvi r onm e nt but i t no w m us t be r e l e ga t e d t o one c om pone nt of a n ove r a l l M A V f l i ght s ys t e m t ha t w i l l t a ke da t a f r om a l l t h r e e s ys t e m s i ne r t i a l vi s i on, a nd G P S a nd de t e r m i ne t he ne c e s s a r y s e r vo c om m a nds t o c ont i nue i t s m i s s i on. T he c om pl e xi t y o f t hi s s ys t e m w i l l be f a r be yond t ha t of t he M A V 128. T he hor i z on t r a c ki ng s ys t e m t he f i r s t c om pone nt of t he M A V v i s i on pr oc e s s i ng s ys t e m de ve l ope d, c oul d s uc c e s s f ul l y r un on a 1 G H z pr oc e s s or onl y t hr ough t he us e of a ha r dw a r e f r a m e gr a bbe r A D S P o r a pr oc e s s or w i t h s pe c i a l f unc t i ons f or i m a ge pr oc e s s i ng, w oul d m os t l i ke l y ha ve t o be us e d t o ha ve a ny c ha nc e of r unni ng t h e hor i z on t r a c ki ng s ys t e m onboa r d. E ve n by r e duc i ng t he c om pl e xi t y of t he a l gor i t hm w e w oul d s t i l l f a c e i m m e ns e c ha l l e nge s i n de ve l opi ng a t r a di t i ona l pr oc e s s i ng s ys t e m t ha t c oul d a c c om pl i s h t hi s gi ve n t he s i z e w e i ght a nd pow e r l i m i t a t i ons o f a M A V pl a t f or m F ur t he r m or e ha vi ng ho r i z on t r a c ki ng onboa r d w i l l a l l ow us t o e a s i l y a ugm e nt t he i ne r t i a l ba s e d s t a bi l i z a t i on s ys t e m w i t h vi s i on a nd p r oduc e a m o r e r obus t c ont r ol l e r S i m i l a r l y, t he r e a r e f a r m or e c om pl e x vi s i on pr oc e s s e s t ha t a r e j us t now c om i ng on l i ne t h a t m us t e ve nt ua l l y be m ove d onboa r d

PAGE 83

72 i f w e a r e t o s uc c e e d i n de ve l opi ng a n a ut onom ous M A V c a pa bl e of ur ba n ope r a t i ons [ 25, 26] F or t una t e l y, w e ha ve m or e opt i ons a va i l a bl e t o us t he n t he t r a di t i ona l pr oc e s s i ng s ys t e m s W hi l e i ni t i a l l y t he P L D m a r ke t w a s f oc us e d on r e pl a c i ng s i m pl e l ogi c s ys t e m s w i t h a s i ngl e de vi c e t he a dva nc e s i n F i e l d P r og r a m m a bl e G a t e A r r a ys ( F P G A s ) ove r t he pa s t f e w ye a r s ha s l e d t o g r ow i ng i n t e r e s t ove r t he a bi l i t y of t he de vi c e s t o f unc t i on a s c us t om de s i gns i n m ul t i pl e a ppl i c a t i ons T he ne w e s t de vi c e s a r e s o a dva nc e d t ha t t he y a r e now be i ng us e d t o not onl y a ugm e nt D S P s ys t e m s but a l s o r e pl a c e t he m e nt i r e l y B e c a us e of t hi s F P G A ba s e d s ys t e m on a c hi p ( S oC ) t e c hnol ogy i s be i ng us i ng i n m a ny a ppl i c a t i ons i nc l udi ng vi s i on p r oc e s s i ng. J us t r e c e nt l y, F P G A s ha ve be e n us e d i n S ony hum a noi d r obot s a s a m e a ns f o r s uppor t i ng s t e r e o vi s i on by r unni ng t he ne c e s s a r y a l gor i t hm s t o c om bi ne t he t w o c a m e r a s i gna l s f or pr oc e s s i ng [ 27] T he a bi l i t y t o de ve l op a c us t om di gi t a l s ys t e m i n a n F P G A a s w e l l a s ha ve t hi s de s i gn pe r f or m m ul t i p l e ope r a t i ons i n a s i ngl e c l oc k c yc l e m a ke F P G A s w e l l s ui t e d f or pe r f o r m i ng t he i nt e ns i ve pi xe l ope r a t i ons ne c e s s a r y i n m os t vi s i on p r oc e s s i ng a ppl i c a t i ons T he r e f or e t he ne xt s t a ge of f l i ght c om put e r de ve l o pm e nt w i l l be he a vi l y f oc us e d on de s i gni ng a n F P G A ba s e d s ys t e m c a pa bl e of pe r f or m i ng hor i z on t r a c ki ng. I t w i l l a l s o ha ve t he pr oc e s s i ng pow e r a nd c a pa c i t y t o ha ndl e m or e t he n j us t t he vi s i on ba s e d s t a bi l i t y s ys t e m s o t ha t ot he r c om pone nt s of t he gr ound s t a t i on vi s i on s ys t e m s m a y be por t e d t o t he s ys t e m T he s a m e c ha l l e nge s f a c e d i n t he de ve l opm e nt o f t he M A V 128 i n t he pa s t ye a r a nd a ha l f w i l l s t i l l be m a j or f a c t or s i n t he de ve l opm e nt o f t hi s ne w s ys t e m but e m e r gi ng t e c hnol ogy de ve l ope d by t he i ndus t r y a l l o w e d us t o m ove be yond t hos e c ha l l e nge s a nd de ve l op a f l i ght a vi oni c s s ys t e m c a pa bl e of a ut onom ous c ont r o l f o r s uc h a

PAGE 84

73 s m a l l a i r c r a f t A s s uc h, ne w de ve l opm e nt s c om i ng i n t he ne xt ye a r or s o w i l l a l s o a l l ow us t o a dd vi s i on p r oc e s s i ng c a pa bi l i t i e s t o t he s ys t e m a nd m ove us c l os e r t o our goa l of de ve l opi ng a n a ut onom ous M A V c a pa bl e of u r ba n ope r a t i ons

PAGE 85

74 L I S T O F R E F E R E N C E S [ 1] J M M c M i c ha e l a nd C ol M S F r a nc i s M i c r o A i r V e hi c l e s T ow a r d a N e w D i m e ns i on i n F l i ght W or l d W i de W e b, ht t p: / / w w w da r pa m i l / t t o/ m a v/ m a v_a uvs i ht m l A u gus t 1997 [ 2] P G I f j u, S E t t i nge r D A J e nki ns Y L i a n W S hy, a nd M R W a s z a k, F l e xi bl e w i ng ba s e d M i c r o A i r V e hi c l e s 40 t h A I A A A e r os pac e Sc i e nc e s M e e t i ng R e no, N V A I A A 2002 0705 [ 3] P G I f j u, S E t t i nge r D A J e nki ns a nd L M a r t i n e z C om pos i t e M a t e r i a l s f o r M i c r o A i r V e hi c l e s S A M P E J our nal vol 37, no. 4, pp. 7 13, J ul y/ A ugus t 2001 [ 4] S E t t i nge r M C N e c hyba P G I f j u, a nd M W a s z a k, V i s i on gui de d F l i ght S t a bi l i t y a nd C ont r ol f o r M i c r o A i r V e hi c l e s P r oc I E E E I n t C onf on I nt e l l i ge nc e R obot s and Sy s t e m s vol 3, pp. 2134 40, 2002 [ 5] S E t t i nge r M C N e c hyba P G I f j u a nd M W a s z a k, V i s i on gui de d F l i ght S t a bi l i t y a nd C ont r ol f o r M i c r o A i r V e hi c l e s A dv anc e d R obot i c s vol 17 no 7 pp. 617 40 2003 [ 6] A K ur di l a V i s i on B a s e d C ont r ol of A gi l e A ut o nom ous M i c r o A i r V e hi c l e s a nd S m a l l U A V s i n U r ba n E nvi r onm e nt s P r oj e c t O ve r vi e w W or l d W i de W e b, ht t p: / / w w w m i l uf l e du/ m a v/ pr e s e nt a t i ons / ki c kof f _m t g/ ove r vi e w pdf A V C A A F K i c kof f M e e t i ng, U F G E R C 27 O c t obe r 2003 [ 7] L A r m e s t o, S C hr ous t M V i nc z e a nd J T or ne r o M ul t i r a t e F us i on w i t h V i s i on a nd I ne r t i a l S e ns or s P r oc I E E E I nt C onf on R o bot i c s and A ut om at i on vol 1 pp. 193 99 2004 [ 8] 3D M G U s e r M a nua l M i c r os t r a i n, I nc W or l d W i de W e b, ht t p: / / w w w m i c r os t r a i n c om / us e r m a nua l s / 3D M G us e r m a nua l pdf A p r i l 2003 [ 9] S w i f t A 2 G P S R e c e i ve r P r oduc t S pe c i f i c a t i on, A xi om N a vi ga t i on, I nc C os t a M e s a C A 2002 [ 10] J G r z yw na S K a now i t z P I f j u a nd M N e c hyba I nt e gr a t i ng G P S w i t h M i c r o A i r V e hi c l e s W or l d W i de W e b, h t t p: / / w w w m i l uf l e du/ ~ num be r 9/ m a v/ D e c e m be r 2002

PAGE 86

75 [ 11] G P S R e c e i ve r M e s s a ge S e t S pe c i f i c a t i on, A xi om N a vi ga t i on I nc C os t a M e s a C A 2002 [ 12] C om pa c t R F U s e r s G ui de M i c r oha r d C or p C a l ga r y, A B C a na da 2003 [ 13] M H X 910/ 2400 U s e r s G ui de M i c r oha r d C or p. C a l ga r y, A B C a na da 2003 [ 14] A T m e ga 128( L ) D a t a s he e t A t m e l C or p S a n J os e C A 2004 [ 15] J W G r z yw na D M a c A r t hur J P l e w a nd M C N e c hyba E va l ua t i on of a M A V I ne r t i a l ba s e d F l i ght S t a bi l i t y S ys t e m us i ng V i s i on F e e dba c k, a c c e pt e d t o A I A A I nt e l l i ge nt S ys t e m s C onf e r e nc e C hi c a go, I L S e pt e m be r 2004. [ 16] J W G r z yw na A J a i n, J P l e w a nd M C N e c hyba R a pi d D e ve l opm e nt of V i s i on B a s e d C ont r ol f or M A V s t hr ough a V i r t ua l F l i ght T e s t be d, s ubm i t t e d t o I E E E I nt C onf on R obot i c s a nd A ut om a t i on B a r c e l ona S pa i n, A pr i l 2005. [ 17] J W G r z yw na A F l i ght T e s t be d W i t h V i r t ua l E n vi r onm e nt C a pa bi l i t i e s f or D e ve l opi ng A ut onom ous M i c r o A i r V e hi c l e s M a s t e r s T he s i s U ni ve r s i t y of F l or i da 2004 [ 18] S pe c i f i c a t i ons f or G P S R e c e i ve r M ode l G H 80 F u r uno E l e c t r i c C o. L t d. S ys t e m P r oduc t s D i vi s i on, C a m a s W A 2002 [ 19] A C 4490 D a t a s he e t A e r oc om m L e ne xa K S 2003 [ 20] A C 4490 U s e r s G ui de V 1. 7 A e r oc om m L e ne xa K S 2004 [ 21] S J ung, K L e e P A B a r ns w e l l P G I f j u J W G r z yw na J P l e w A J a i n, a nd M C N e c hyba V i s i on ba s e d C ont r ol f or a M i c r o A i r V e hi c l e : P a r t 1 : T e s t be d, s ubm i t t e d t o A I A A C onf f or G ui da nc e N a vi ga t i o n, a nd C ont r o l P r ovi de nc e R I A ugus t 2004. [ 22] J K e hoe J W G r z yw na R S C a us e y, J P l e w M A bdul r a hi m M C N e c hyba a nd R L i nd, W a ypoi nt N a vi ga t i on f o r a M i c r o A i r V e hi c l e us i ng V i s i on B a s e d A t t i t ude E s t i m a t i on s ubm i t t e d t o E ur ope a n M i c r o A i r V e hi c l e C onf B r a uns c hw e i g, G e r m a ny, J ul y 2004 [ 23] K e s t r e l A ut opi l ot 1 45 D e s c r i pt i on, P r oc e r us T e c hnol ogi e s P r ovo U T 2004 [ 24] a ut opi l ot : D o i t your s e l f U A V W or l d W i de W e b, ht t p: / / a ut opi l ot s our c e f o r ge ne t / N ove m be r 2004 [ 25] S T odor ovi c a nd M C N e c hyba A V i s i on S ys t e m F or I nt e l l i ge nt M i s s i on P r of i l e s of M i c r o A i r V e hi c l e s a c c e pt e d by I E E E T r ans O n V e hi c ul ar T e c hnol ogy I n pr e s s

PAGE 87

76 [ 26] S T odor ovi c M C N e c hyba a nd P G I f j u, S ky / G r ound M ode l i ng f o r A ut onom ous M A V s P r oc I E E E I nt C onf R obot i c s and A ut om at i on vol 1 pp 1422 7, S e pt e m be r 2003 [ 27] K S a be M F ukuc hi J S G ut m a nn, T O ha s hi K K a w a m ot o, a nd T Y os hi ga ha r a O bs t a c l e A voi da nc e a nd P a t h P l a nn i ng f or H um a noi d R obot s U s i ng S t e r e o V i s i on, P r oc I E E E I nt C onf on I n t e l l i ge n c e R obot s and Sy s t e m s vol 4 pp. 3488 93 2004

PAGE 88

77 B I O G R A P H I C A L S K E T C H J a s on P l e w w a s bor n i n B e df or d I N i n 1979 H e f i r s t be ga n r e s e a r c h i nt o r obot i c s w he n e nt e r i ng hi gh s c hool i n P a l m B a y, F L U pon gr a dua t i ng hi gh s c hool i n 1998, J a s on be ga n a t t e ndi ng t he U ni ve r s i t y o f F l or i da w he r e h e s t a r t e d w or ki n g a t t he M a c hi ne I nt e l l i ge nc e L a b. D ur i ng t hi s t i m e he ha s a l s o w o r ke d t hr ough s um m e r i nt e r ns hi ps a t bot h P C om I nc a nd C e nt e r poi nt B r oa dba nd T e c hnol ogi e s I nc J a s on l a t e r be ga n t o w or k a t P r i or i a R obot i c s I nc a s m a l l s t a r t up c om pa ny l oc a t e d i n G a i ne s vi l l e F L I n 2003, he g r a dua t e d f r o m U F w i t h B a c he l or o f S c i e nc e de gr e e s i n bot h c om put e r a nd e l e c t r i c a l e ngi ne e r i ng. H e t he n be ga n pu r s i ng a M a s t e r of S c i e nc e i n e l e c t r i c a l e ngi ne e r i ng, us i ng t he M A V r e s e a r c h he w a s c onduc t i ng a s t he ba s i s f or t hi s t h e s i s