`
`FOR MOBILE o B01-5?
`
`THEORY AND APPLICATION
`
`SilverStar Exhibit 1016
`SilverStar Exhibit 1016
`
`

`

`I591! Insaneawea
`
`SilverStar Exhibit 1016 - 2
`SilverStar Exhibit 1016 - 2
`
`

`

`Sensors
`
`for Mobile
`
`Robots
`
`Theory and Application
`
`HR. Everett
`Na val Command, Contra! and
`Ocean Surveillance Center
`San Diego. California
`
`A K Peters, Ltd.
`Wellesley, Massachusetts
`
`SilverStar Exhibit 1016 - 3
`SilverStar ExhibitTo16 - é
`
`

`

`Editorial. Sales. and Customer Service Office
`
`A K Peters, Ltd.
`289 Linden Street
`Wellesley, MA 02181
`
`Copyright © 1995 by A K Peters. Ltd.
`
`All rights reserved. No part ot the material protected by this copyright notice may
`be reproduced or utilized in any torm. electronic or mechanical. inciuding photo-
`copying. recording, or by any information storage and retrieval system. without
`written permission from the copyright owner.
`
`Library of Congress Cataloging-in-Publicatton Data
`
`Everett. H. FL. 1949-
`Sensors for mobile robots : theory and application I H.Ft. Everett.
`p. cm.
`tnciudes bibliographical references and index.
`ISBN 1-56881-048-2
`1. Mobile robots. 2. Robots—Control systems.
`TJ211.415.E83 1995
`629.8 ' 92—dc20
`
`I.
`
`'l'rtle.
`
`95-17178
`Cl P
`
`Many oi the designations used by manufacturers and sellers to distinguish their products are
`ciaimed as trademarks. Where designations appear in this book and A K Peters was aware of the
`trademark claim. the designations have been printed in italics. Where designations have not been
`provided. every effort has been made to ensure accurate presentation oi product names and
`specifics.
`
`Principle illustrator: Todd Ashley Everett
`
`Printed in the United States of America
`999897969510987654321
`
`SilverStar Exhibit 1016 - 4
`_Si|vérStar Exhibit 1016 - 4
`
`

`

`Table of Contents
`
`FOREW0RD.-................ ................xiii
`
`
`PREFACE............
`.....
`.....
`
`“.....xv
`
`ACKNOWLEDGMENTS ..........................................................................................................xvii
`
`I. INTRODUCTION .................................................................................................... .. ................. l
`
`
`
`1.] DESIGN CONSIDERATIONS ........
`1.2 THE ROBOTS...
`
`l.2.| WALTER(I965AI967J
`
`1.2.2 CRAWLER I “9664968)...
`
`
`1.2.3 CRAWLER lI (1968—l97l J ....................................................
`
`1.2.4 RDBARTI (IQSII- [935}
`l. 2.5 ROBAR'E' II “982— J“
`
`l2.-6MODBOT(1990].
`l.2.7' USMC TeleUpcratcd Vehicle(l935- I989J...
`
`l.2.I5 MDARS lnteriofl “#39— J ...............................
`..
`l.2.'9 Surrogate Tcleopcrated Vehicle (1990- 1993)
`..25
`
`..............
`..28
`1.2.10 ROBARTIII (1992-)......
`I.2.l1 MDARS Exterior (£9949...
`
`
`
`
`2. DEAD RECKONING
`
`
`2.] ODOMETR‘Ir SENSORS.......
`
`2. I. | Potentiometers
`2 |. 2 Synchros and Resulvera.
`
`2.l. 3 Optical Encoders"
`2. 2 DOPPLER AND {HERriai. NAVIGATION
`
`2.2. | Doppler Navigaiinn...
`
`2.2 2 lnerliai Navigation”
`2.3 TYPICAL MOBILITY CONFIGURATIONS
`
`2.3.I Differential Steering...
`
`2.3,2 Ackcrman Steering.
`
`2.3.3 Synchm Drive"...
`
`
`2. 3 4 Tricycle Drive.
`2.35 Omni-Directional Driviz...
`
`24 INl‘ERNAL POSITION ERROR CORRECTION
`
`2.5 REFERENCES
`3. TACTIIJEJ AND PROX‘IMITY SENSING .....u....69
`
`3.|
`
`'I‘ACI‘ILESENSORQ
`3.1.1 Taclile Peelers
`1L2 Tactile Bumpeis
`1L3 Disnibulul Surface Arrays
`
`
`
`
`SilverStar Exhibit 1016 - 5
`SilverStar Exhibit 1016 - 5
`
`

`

`vi
`
`Sensors for Mobile Robots
`
`3.2 PROXIMITY SENSORS .....
`
`3.2.] Magnetic Proximity Sensors
`
`3.2.2 Inductive Proximity Sensors......
`
`3.2.3 Capacitive Proximity Sensors....
`
`3.2.4 Ultrasonic Proximityr Sensors.
`3.2.5 Microwave Proximity Sensors...
`
`3.2.6 Optical Proximity Sensors ..
`3.3 REFERENCES
`
`
`4. TRMNGULATION RANGING .................................................................
`
`................103
`
`4| STEREO DISPARITY...
`....lflfi
`.109
`4.1.] .‘iPL Stereo Vision .....
`4 I. 2 David Sarnoff Stereo Vision
`.III
`
`.
`I I4
`42 ACTIVE TRIANG ULATION.
`
`4.2. l Hamamalsu Rangefinder Chip set.
`.I If:
`
`4 2. 2 Draper Laboratory Rangefinder .........
`...I 1’!
`
`4.2.3 Quanric Ranging System .....
`I I9
`.121
`4.3 ACTIVE STEREOSCOPIC ..
`
`....122
`4. 3 ] HERMIES
`
`43 2 Dual—Aperture 3DRzingeSensor... {24
`44 STRUCTURE-1D LIGHT"
`.125
`
`44 | TRC SirobedLight TnangulatlonSystem...
`[27
`4.5 KNOWN TARGET SIZE.
`I28
`4.5. I NAMCO Lasemela ScanningLasei'Sensor
`. 129
`4.6 OPTICAL FLOW
`.
`1 3i
`4.6. I NIST PEISSIVE Ranging and CoIIISIon Avoidance.
`I33
`
`4.6. 2 David Sarnoff PaS5ive Viston .....
`.133
`
`4.? REFERENCES .................................
`I34
`
`
`
`
`5. TIME OFFLIGHT. .................... 139
`
`
`
`MI
`5. I ULTRASONIC TOF SYSTEMS.
`I4l
`5. I. 1 National Semiconductor‘5 LMISI2Ultrasonic Transceiver
`
`143
`5.1.2 Massa Products Ultrasonic Ranging Module Subsystems
`
`144
`5.I.3 Polaroid Uiirasonic Ranging Modules....................
`
`148
`5.1.4 Cybermoiion (CA-2 Collision Avoidance System
`I50
`5.2 LASER—BASED TOF SYSTEMS
`
`
`I50
`5.2. I Schwartz Electra—Optics Laser Range 1
`
`[53
`5.22 RIEGL Laser Measurement Syslem5...
`S2.3 Odetics FasI Frame Rate 3~D Laser ImagingSysiem.16]
`
`
`
`5. 3 REFERENCES...
`165
`5.2.4 RVSI LungOptical Ranging and Detection System
`I62
`
`6. PHASE~SHIET MEASUREMENT AND FREQUENCY MODULATION“..................... 169
`
`(ml PHASE—SHIFT MEASUREMENT
`
`6.1.1 ERIM 3-D Vision Systems..
`
`6.1.2 Perceprron LASAR ..............
`
`6.1.3 Odetics Scanning Laser Imaging System.
`
`6. | .4 Sandia Scanncrlcss Range Imuger
`
`6.1.5 ESP Optical Ranging System ......
`6.1.6 Acuity Research AccuRange 3(300 .
`
`
`I69
`I74
`IT!
`['38
`180
`183
`l 85
`
`
`SilverStar Exhibit 1016 - 6
`SilverStar' Exhibit 1016 - 6
`
`

`

`Tahie of Contents
`
`vii
`
`
`6.1.7 TRC Light Direction and Ranging System..........
`............ IS"!
`
`.
`............ ISB
`6. 2 FREQUENCY MODULATION.
`..
`6.2. l VRSS Automotive Collision Avmdance Radar.............................................................. 199
`
`62.2 VORAD Vehicle Detection and Driver Alert System.
`. 19]
`623 Safety First Systems Vehicular Obstacie Detection and Warning System...
`.193
`6.2 4 Millilech Millimeter Wave Radar...
`. [94
`6.3 REFERENCES
`
`
`
`7. OTHER RANGING TECHNIQUES.....
`
`
` 1 l INTERFEROMETRY. ...
`.201
`7 l. I CLS Coordinate MeasiiringSysierii
`
`7.2 RANGE FROM FOCUS...
`.202
`.203
`
`7.2.I Honeywell Autofouus Sysiems...
`
`......206
`7.2.2 Associates and Ferren Swept-Focus Ranging.....................
`
`....EIO
`7.2.3.1?1. Range‘from-Focus System
`.21]
`13 RETURN SIGNAL INTENSITY
`
`....212
`7.3.l Programmable Near-Infrared Proximity Sensor ......
`
`....2l5
`7.3.2 Australian National University Rangefinder.....
`.216
`7.3.3 MIT Near—Inhaled Ranging System ..............
`
`.216
`I14 Honeywell Displaced—Sensor Ranging Unit .
`7.4 REFERENCES...
`...................... 2|?
`
`
`8. ACOUSTICAL ENERGY...22!
`
`
`
`8. | APPLICATIONS.
`...................... 224
`3. 2 PERFORMANCE FACTORS
`225
`
`225
`8.2.1 Atmospheric Attenuation ..
`
`227
`8.2.2 Target. Reflectivity .....
`
`3.2.3 Air Turbulence
`232
`233
`
`8.2.4 Temperature
`.......234
`8.2.5 Beam Geometry ............
`
`...239
`8. 2.6 Noise ................................
`240
`
`827 SystemSpecific Anomalies...
`....242
`8 3 CHOOSING AN OPERA]ING FREQUENCY
`
`....242
`3.4 SENSOR SELECTION CASE STUDY.
`
`5.5 REFERENCES
`....244
`.
`
`
`9. ELECTROMAGNETIC ENERGY........
`
`.............................................................249
`
`9.] OPTICAL ENERGY......
`....252
`....253
`9. l. l Eiectro--Opticai Sources...
`
`
`....... 258
`9.1.2 Performance Factors...
`.
`....262
`9 | 3 Choosing an Operating Wavelength
`
`9.-7 MICROWAVE RADAR"
`...263
`...264
`9.2.] Appiicationh‘
`
`9.2.2 Performance FaLlors ......
`264
`
`26‘?
`9.3 MILLIMETERWAVE RADAR ..
`268
`
`93. I Applications,
`“.269
`9.3 .2 Perle-finance Factors
`274
`9.3.3 Choosing anOperating Frequency"
`
`
`274
`9. 4 REFERENCES...
`... ..
`
`
`SilverStar Exhibit 1016 - 7
`SiliierStar Exhibit 1016 - 7—
`
`

`

`viii
`
`Sensors for Mobile Robot:-
`
`lfl. COLLISION AVOIDANCE...........................
`
`.................................279
`
`
`279
`10.1 NAVlGATIONAJ. CONTROL STRATEGIES ..........
`....280
`£111.] Reactive Control...................................
`
`[0. l .2 Represenialional World Modeling.
`”.290
`
`10.1.3 Combined Approach .....................
`..,.295
`
`[0.2 EXTERIOR APPLICATION C0 SIDERATIONS...
`....299
`
`10.3 NAVIGATIONAL RErREFERENCl'NG .............
`....301
`
`....302
`10.4 REFERENCES.........................,..
`
`ll. GUIDEPATH FOLLOWING ................................“.305
`
`
`
`11.] WIRE GUIDED ..............................................................
`l 1.2 OPTICAL STRIPE
`
`
`l 1.2.] ModBoI Optical Stripe Tracker ............
`
`1122 UN Stimulated Emission
`l L3 MAGNETIC TAPE ............................
`
`1 1.3.1 Macome Magnetic Stripe Follower...
`11.3.2 Apogee Magnetic SLrip-e Follower....
`1 1.3.3 JMH-loneyweil Magnetic Lateral Guidance ystcm
`
`1 L4 HEAT AND ODOR SENSING ....................
`|
`|.5 INTERMITFENT—PATH NAVIGATION
`
`
`I I.5.| MDARS. Interior Hybrid Navigalinrl ...... ..
`”.52 Free Ranging On Grid ,,,,,
`
`1 L6 REFERENCES .........................
`
`12. MAGNETIC CDMPASSES...................................................................................... 327
`
`[ill MECHANICAL MAGNE'I'IC COMPASSES
`
`
`12. l ,1 Dinsmorc Starguide Magnetic Compass...
`|2.2 FLUXGATE COMPASSES
`
`[2.2.1 szco Fluxgale Compasses
`
`|2_2.2 Watson Gyro Compass
`
`
`|2.2.3 KVl-I Fluxgale Compabseb
`
`12.2.4 Applied Physics Systems Minialure Orientation Sensor.
`
`[23 MAGNETOTNDUCTIVE MAGNET‘OMETERS
`
`12.3.1 Precision Navigation TCM Magneloinductivc Compass..._
`
`
`|2.4 HALL-EFFECT COMPASSES ..............
`I25 MAGNETORESISTIVE COMPASSES,
`
`|2.S.l Philips AMR Compass
`
`12.5.2 Space Electronics AMR Compass“
`12.5.3 Honeywell HMR Series Smart Digital Magnetomeler
`12.6 MAGNETOELASTIC COMPASSES
`
`[2.7 REFERENCES ...................................................................................................................... 357
`
`13. GYROSCUPESun......361
`
`13. | MECHANICAL GYROSCOPES ...........
`
`l1]. 1 Space—Stable Gyroscopes _
`
`13.1.2 Gyrocompasses ...............................
`
`13.1.3 RateGyros
`
`
`SilverStar Exhibit 1016 - 8
`SilverStar Exhibit 10—16 - 8
`
`

`

`Table of Contents
`
`ix
`
`
`....371
`13.2 OPTICAL GYROSCOPES .............
`
`
`373
`I111 Active Ring-Laser Gyros ......
`.....330
`|3.2.2 Passive Ring Resonator Gyros ..............................
`
`..SBI
`|3.2.3 Open—Loop Interferometric Fiber-Optic Gyros .
`
`..387
`1.3.2.4 Closed-me Interferometric Fiber-Optic Gyros...
`
`.389
`13.2.5 Resonant Fiber—Optic Gyros.
`
`13.3 REFERENCES
`....................... 390
`
`14. RF POSITION-LOCATION SYSTEMS
`
`.......
`
`
`
`............395
`
`....395
`[4. l GROUND- BASED RF SYSTEMS...
`395
`14 I. I Loren.
`
`I4. I'2 Kaman SciencesRaciio Freqiiency Navigation Grid .............. 396
`
`[4. | 3 Precision Technology Tracking and Telemetry System
`.398
`[4.1.4 Motorola Mini-RangerFalcon
`
`40]
`141.5 Harris infogeometric System
`.403
`14.2 SATELLITE-BASED SYSTEMS
`.403
`
`14.21 Transit Satellite Navigation System
`.405
`[4.2.2 Navstar Global Positioning System.
`
`I43 REFERENCES
`420
`
`15. ULTRASONIC AND OPTICAL POSITION-LOCATION SYSTEMS...........................423
`
`l5.l ULTRASONIC POSITION-LOCATION SYSTEMS.......
`
`I5. I. I Ultrasonic Transponder Trilateration ........................
`
`l5. [.2 Ultrasonic Signature Matching“
`15.2 OPTICAL POSITION—LOCATION SYSTEMS
`
`15.2.1 CRAWLER I Homing Beacon .........
`
`I521 ROBART ll Recharging Beacon ..
`
`I5.23 Cyberrnotion Docking Beacon ..
`l5.2.4 Hilare.
`..
`
`IS 2.5 NAMCOLes-erneia Scanning Latter Sensor
`
`[5.26 Caterpillar SelfGuided Vehicle.
`.
`
`15.2.7 TRC Beacon Navigation System
`|S.2.8 Intelligent Solutions EZNIIV Position Sensor .....
`
`15.2.9 Imperial College Beacon Navigation System .....
`
`15.2.10 MTI Research CONAC ..............
`
`15.1.1 I MDARS Lateral—Post Sensor...
`
`I53 REFERENCES..........
`
`
`16. WALL, DOORWAY, AND CEILING REFERENCINGKISS
`
`I6] WALL REFERENCTNG .....
`....455
`....455
`16.].I Tactile Wall Referencing...
`
`....458
`I6. I 2Non-Con1actWa‘ll Referencing.
`
`....461
`16.1.3 Wall Following ....................
`....465
`16.2 DOORWAY TRANSIT REFERENCING
`
`....472
`16.3 CEILING REFERENCING
`472
`16.3. l Polarized Optical Heading Reference..........
`
`473
`l6. 3. 2 Georgia Tech Ceiling Referencing System
`474
`
`I633 TRC HelpMaae Ceiling Referencing System...
`....4?6
`I634 MDARS Overhead-Beam Referencing System ..
`
`....477
`
`
`
`SilverStar Exhibit 1016 - 9
`SilverStar Exhibit 1016 - 9
`
`

`

`x
`
`Sensors for Mobile Robots
`
`1?. APPLICATION-SPECIFIC MISSION SENSORS...
`
`17,1 THE SECURITY APPLICATEON ..
`
`
`l'?_|_| Acoustical Detection _
`iT.l.2 Vibration Sensors
`
`FMS Ultrasonic Presence Sensors.
`IT. | .4 Optical Motion Detection
`
`
`l7. LS Passive Infrared Motion Detection.
`[1L6 Microwave Motion De:ection................
`
`I7. If! Video Mminn Detectinn...............,,
`
`17.1.3 intrusion Detection on the Move.
`
`17.1.9 Verification and Assessment
`
`17.2 AUTOMATED INVENTORY ASSESSMENT...
`
`Ill! MDARS Product Assessment System
`|?.3 REFERENCES....._.,............
`.......
`
`
`
`SilverStar Exhibit 1016 - 10
`SilverStar Exhibit 1016 - 10
`
`

`

`Foreword
`
`A robot’s ability to sense its world and change its behavior on that basis is what
`makes a robot an interesting thing to build and a useful artifact when completed.
`Without sensors. robots would be nothing more than fixed automation. going
`through the same repetitive task again and again in a carefully controlled
`environment.
`Such devices certainly have their place and are often the tight
`economic solution. But with good sensors, robots have the potential to do so
`much more. They can operate in unstructured environments and adapt as the
`environment changes around them. They can work in dirty dangerous places
`where there are no humans to keep the world safe for them. They can interact
`with us and with each other to work as parts of teams. They can inspire our
`imaginations and lead us to build devices that not so long ago were purely in the
`realms of fiction.
`
`Sensors are what makes it all possible.
`When it comes right down to it there are two sorts of sensors. There are visual
`sensors. or eyes, and there are non—visual sensors. Lots of books have been
`written about visual sensors and computer vision for robots.
`There is exactly one book devoted to non-visual sensors. This one.
`We tend to be a little vision-centric in our “view" (there we go again...) of the
`world, as for humans vision is the most vivid sensor mechanism. But when we
`look at other animals, and without
`the impediment of introspection, another
`picture (hmmmuJ begins to emerge.
`Insects have two eyes, each with at most
`perhaps [0,000 sensor elements.
`Arachnids have eight eyes, many of them vestigial, some with only a few
`hundred sensor elements, and at most 10.000 again. But insects have lots and lots
`and lots of other sensors. Cockroaches, for example, have 30,000 wind-sensitive
`hairs on their legs, and can sense a change in wind direction and alter the direction
`in which they are scutlling in only 10 milliseconds. That is why you cannot stomp
`on one unless you have it cornered. and on top of that get lucky. The cockroach
`can sense your foot coming and change course much faster than you can change
`where you are aiming. And those 30,000 sensitive hairs represent just one of a
`myriad of specialized sensors on a cockroach. Plus each different insect has many
`varied and often uniquely different sensors. Evolution has become a master at
`producing non-visual sensors.
`As robotics engineers we find it hard to create new sensors, but are all aware
`that in general our robots have a rather impoverished connection to the World.
`More sensors would let us program our
`robots in ways that handled more
`situations, and do better in those situations than they would with fewer sensors.
`Since we cannot easily create new sensors. the next best thing would be to know
`what sensors were already available. Up until this point we have all maintained
`
`
`SilverStar Exhibit 1016 - 11
`SilverStar Exhibit 1016 - 11
`
`

`

`xii
`
`Sensors for Mobile Robots
`
`our own little libraries of sensors in our heads. Now Bart Everett has written
`
`dowri all he had in his own private library and more. Ban‘s robots have always
`stood out as those with Lhc most sensors. because interactive sensing has aJWays
`been a priority for Bart. Now he is sharing his accumulated wisdom with us, and
`robotdorn will be a better place for it. Besides providing us with an expanded
`library. Bart has also done it in a way that everyone interested in robotics can
`understand. He takes us through the elementary physics of each sensor with an
`approach that a computer scientist. an electrical engineer. a mechanical engineer.
`or an industrial engineer can relate to and appreciate. We gain a solid
`understanding of just what each sensor is measuring, and what its limitations will
`be.
`
`So let‘s go build some new robots!
`
`Rodney A. Brooks
`MIT Al Lab
`
`Cambridge. MA
`
`
`SilverStar Eihibit 1016 - 12
`SilverStar Exhibit 1016 - 12
`
`

`

`Preface
`
`in the preparation of this manuscript was to present some
`My underlying goal
`general background on the sensing needs of a mobile system.
`followed by
`sufficient theory of operation and illustrative examples such that the overall result
`is both informative and of practical use. Perhaps the most challenging problem I
`faced early on in this endeavor was how to arrange reams of information on all the
`various sensors into some semblance of logical order. One considered possibility
`was to categorize by class of robot {i.e.. airborne. underwater. indoor. exterior,
`autonomous, teleoperated). Given the emphasis of the book. hOWever. it seemed
`more appropriate to break down the discussion by sensor type.
`In an attempt to bound the problem, I decided to eliminate any treatment of
`airborne or undervvater scenarios and focus instead on interior and exterior land-
`
`based applications. Even so. there was still considerable difficulty associated with
`organizing the flow. For example. at least seven different methods of non-contact
`ranging techniques are known to exist; one of these methods alone (triangulation)
`can be implemented in five different ways. Almost all such ranging systems can
`operate in the acoustical or electromagnetic regions of the energy spectrum; can
`be active or passive; and may have markedly different assigned functions in actual
`deployment.
`After much weighing of alternative strategies, I chose to present the material in
`a manner that to some extent parallels the strategy often employed in robotic
`development.
`The initial
`thrust of most early research efforts in which I
`participated was simply aimed at how to get
`the robot
`to move about
`in a
`controlled and purposeful fashion. Once this hurdle is surmounted. attention can
`be turned to collision avoidance. wherein the system learns not to run into things
`while enroute. The proud builders soon realize the robot can perform admirably
`for
`some finite length of
`time but eventually will get
`lost. whereupon
`developmental focus shifts to navigational referencing. Applications are tacked
`on later. sometimes almost as an afterthought.
`Accordingly, following some general background discussions in Chapter 1, we
`start by taking a look in Chapter 2 at
`the sensors employed in vehicle dead
`reckoning, with a careful analysis of potential error sources.
`Tactile and
`proximity sensors are introduced next
`in Chapter 3. providing a rudimentary
`capability to at least detect potential obstructions in time to stop. Chapters 4
`through 7 provide an overview of the various distance measurement techniques
`available, such as triangulation. time of flight. frequency modulation. phaseashift
`measurement. and interferometry. Related discussion of implementation in the
`acoustical. radio frequency. and electro—optical domains is presented in Chapters 8
`and 9, with a special emphasis on the various factors affecting performance.
`This approach hopefully provides a good foundation for later examining how
`such non—contact
`ranging sensors are employed in specific roles.
`first and
`
`
`SilverStar Exhibit 1016 - 13
`SilVerStar Exhibit 1016 - 13
`
`

`

`xiv
`
`Sensors for Mobile Robots
`
`foremost being in support of collision avoidance [Chapter 10). Navigational
`referencing. the subject of Chapters 11 through 16. is addressed in considerable
`detail as it represents one of the biggest remaining stumbling blocks to successful
`fielding. A few representative samples of application-specific sensors are treated
`in closing in Chapter 17.
`In retrospect, there is considerably less emphasis than I originally intended on
`imagekbased systems, as the subject of machine vision quite obviously could be
`the focus of a book all in itself. And since a number of distinguished individuals
`far better qualified than myself have in fact taken that very objective to task, I
`have purposely limited discussion in this volume. and concentrated instead on
`various alternative (and often less complex) sensing strategies less documented in
`the open literature. Reference is made throughout the text to candidate systems,
`both commercially available and under develoPment, in hopes of complementing
`theory of operation with some practical
`lessons in real-world usage. These
`illustrative examples are called out under separate headings where the discussion
`becomes rather detailed.
`
`I have very much enjoyed the preparation of this manuscript, both in terms of
`what I iearned in the process and the new contacts 1 made with other researchers
`in this exciting field.
`I hope the results as presented here will be useful
`in
`promoting the successful employment of mobile robotic systems
`through
`increased awareness of available supporting technologies.
`
`HR. Everett
`
`San Diego. CA
`
`
`SilverStar Exhibit 1016 - 14
`SilverStar Exhibit 1016 - 14
`
`

`

`Acknowledgments
`
`A number of people have assisted me in my educational and research endeavors
`over the years and collectively contributed to making this book a reality.
`I would
`like to express my heart-felt appreciation to:
`
`My uncles. Gene Everett and Joe Hickey. who introduced me to electronics at
`an early age.
`
`My high school geometry teacher, Mrs. Nell Dear. for providing discipline,
`inspiration, and the mathematical foundation upon which I was to build.
`
`Professor Robert Newton, my thesis advisor at the Naval Postgraduate School,
`who made it possible for me to pursue a rather unonhodox topic in the field of
`mobile robotics.
`
`Vice Admiral Earl B. Fowler. USN (Ret.) for creating a robotics program
`office within the Naval Sea Systems Command. and giving me a job after
`graduate sc hoo].
`
`Dr. Anita Flynn of MIT for all the late nights and weekends we spent hacking
`code and building our own sensors in my basement in Virginia.
`
`Gary Gilbt'eath of the Naval Command Control and Ocean Surveillance Center
`for transfomting ROBART 11 into a truly intelligent machine.
`
`My son, Todd Everett, for his tremendous help in generating all the graphics
`used in the figures.
`
`All those people kind enough to review this manuscript in the various Stages of
`its completion, offering helpful insights on how best to present the material: Ron
`Arkin, Johann Borenstein. Fernando Figueroa. Anita Flynn, Doug Gage. Bob
`Garwood, Tracy Heath. Susan Bower, Robin Laird. Richard Langley, Richard
`Lao. Larry Mathies. and Hon Nguyen.
`
`In addition, portions of the material presented in Chapters 4 through 7 were
`previously published in Sensors and later Robotics and Autonomous Systems
`magazines. and updated in this book with their kind permissions.
`
`Silver—Star ExhibiiTtHé - 15
`SilverStar Exhibit 1016 - 15
`
`

`

`1 I
`
`ntroduction
`
`The past several years have brought about a tremendous rise in the envisioned
`potential of robotic systems, along with a significant increase in the number of
`proposed applications. Well-touted benefits
`typically associated with the
`installation of fixed-location industrial robots are improved effectiveness, higher
`quality. reductions in manpower. as well as greater efficiency, reliability. and cost
`savings. Additional drivers include the ability to perform tasks of which humans
`are incapable, and the removal of humans from demeaning or dangerous
`scenarios.
`
`range of
`suggested an additional
`The concept of mobility has always
`applications beyond that of the typical factory floor, where free-roaming robots
`move about with an added versatility fostering even greater returns.
`Early
`developmental efforts introduced potential systems for fighting fires, handling
`ammunition. transporting materials. and patrolling warehouses and storage areas.
`to name but a few. Most of the resulting prototypes met with unexpected
`difficulty, primarily due to an insufficient supporting technology base. Even
`today, after decades of extensive research and development.
`the successful
`application of mobile robots remains for the most part an elusive dream, with only
`a small handful of fielded systems up and running.
`While a number of technological hurdles have impeded progress, the three
`generally regarded as having the greatest impact are;
`l) computational resources,
`2) communications, and 3) sensors. The first two areas have been addressed for a
`variety of commercial reasons with remarkable progress.
`In just a little over 10
`years we have transitioned from 6502- and ZSO—based personal computers running
`under C/PM with a maximum 64—kilobyte address space,
`to Pentium-based
`systems running at 90 MHZ and addressing up to 32 megabytes of memory. The
`recent surge in popularity of laptop computers has provided an extra impetus. with
`special emphasis on reduced power consumption and extended battery life.
`Wireless local area networks and spread-spectrum technology have likewise
`advanced in kind. to the point where there are now a number of vendors offering
`fullvduplex Ethernet-compatible high-speed datalinks with ranges of several
`miles.
`
`
`SilverStar Exhibit 1016 - 16
`SilverVStar Exhibit 1016—41—6—
`
`

`

`2
`
`Sensors for Mobile Robes
`
`The third category of sensors now stands somewhat alone as the most
`significant technical challenge still facing developers, due primarily to a lack of
`high-volume applications. While there has indeed been some carry-over sensor
`technology from advances in flexible automation for manufacturing, it has fallen
`far short of the explosive growth seen in the computer and communications
`industries.
`Successful adaptation of what progress has been made is further
`hampered by the highly unstructured nature of a mobile robot’s operating
`environment.
`Industrial process—control systems Used in repetitive manufacturing
`scenarios,
`in contrast,
`rely on carefully placed sensors that exploit
`the target
`characteristics.
`Background conditions are arranged to provide minimal
`interference, and often aid in the detection process by purposely increasing the on-
`of differential or contrast. Unfortunately. such optimized configuration control is
`usually no longer possible once mobility is introduced as a factor in the equation.
`Consider for example the issue of collision avoidance:
`any mobile robot
`intended for realrworld operation must be capable of moving around without
`running into surrounding obstructions.
`In practice, however,
`the nature and
`orientation of obstacles are not known with any certainty; the system must be
`capable of detecting a wide variety of target surfaces under varying angles of
`incidence. Control 01" background and ambient conditions may not be possible. A
`priori information regarding the relative positions. orientations. and nature of
`objects within the sensor‘s field of view becomes very difficult to supply.
`The situation only worsens when the operating environment is taken outdoors.
`for a number of reasons. To begin with, problems of scale introduce a need for
`additional range capability that significantly adds to system complexity and cost.
`While an indoor collision avoidance system may need to see only 4 to 6 feet in
`front of the robot. for example, exterior scenarios typically require effective
`coverage over a 20- to 30-foot span. sometimes more.
`In addition, the outdoor
`environment often poses additional complicating hazards to safe navigation (i.e.,
`terrain traversabilty. oncoming traffic. atmospheric obscurants)
`that demand
`appropriate engineering solutions not even addressed on interior systems.
`On the positive side. worldwide interest in a rapidly expanding field known as
`intelligent vehicle highway systems- (ll/HS) has already created a huge potential
`market for sensors to address many of these problems as faced by the automotive
`industry (Catling, 1994). Lower-volume autonomous mobile robot applications
`are sure to benefit from the inevitable spin-off technologies that have already
`begun to emerge in the form of lowecost laser and millimeter-wave systems. for
`example. Many of these new and innovative products will be presented as
`illustrative examples in the following chapters, in hopes of further stimulating this
`technologyetransfer process.
`
`1.1 Design Considerations
`
`The problems confronting most mobile robotic development efforts arise directly
`from the inherent need to interact with the physical objects and entities in the
`
`
`SilverStar Exhibit 1016 - 17
`SilverStar Exhibit 1016 - 17
`
`

`

`Chapter 1
`
`Introduction
`
`3
`
`environment. The platform must be able to navigate from a known position to a
`desired new location and orientation, avoiding any contact with fixed or moving
`objects while en route. There has been quite a tendency in early developmental
`efforts to oversimplify these issues and assume the natural growth of technology
`would provide the needed answers. While such solutions will ultimately come to
`pass.
`it
`is
`important
`to pace the evolution of the platform with a parallel
`development of the needed collision avoidance and navigation technologies.
`Fundamental in this regard are the required sensors with which to acquire high-
`resolution data describing the robot's physical surroundings in a timely yet
`practical
`fashion.
`and in keeping with the
`limited onboard energy and
`computational resources of a mobile vehicle. General considerations for such
`sensors are summarized below:
`
`o
`
`Fieid of view — Should be wide enough with sufficient depth of field to
`suit the application.
`0 Range capability — The minimum range of detection, as well as the
`maximum effective range, must be appropriate for the intended use of the
`sensor.
`
`O
`
`- Accuracy and resolution — Both must be in keeping with the needs of the
`given task.
`0 Ability to detect all objects in environment — Objects can absorb emitted
`energy; target surfaces can be specular as opposed to diffuse reflectors;
`ambient conditions and noise can interfere with the sensing process.
`real—
`I Rani-time operation — The update frequency must provide rapid,
`time data at a rate commensurate with the platform's speed of advance
`( and take into account the veloeity of other approaching vehicles).
`0 Concise. easy to interpret data —— The output format should be realistic
`front the standpoint of processing requirements; too much data can be as
`meaningless as not enough; some degree of preprocessing and analysis is
`required to provide output only when action is required.
`G Redundancy — The system should provide graceful degradation and not
`become incapacitated due to the loss of a sensing element; a multimodal
`capability would be desirable to ensure detection of all targets. as well as
`to increase the confidence level of the output.
`Simplicittv —- The system should be low-cost and modular to allow for
`easy maintenance and evolutionary upgrades, not hardware-specific.
`Power consumption -— The power requirements should be minimal
`keeping with the limited resources on board a mobile vehicle.
`Size —- The physical size and weight of the system should be practical
`with regard to the intended vehicle.
`
`II
`
`I
`
`in
`
`The various issues associated with sensor design, selection. and/or integration
`are complex and interwoven. and not easily conveyed from a purely theoretical
`perspective only. Actual device characterization in the form of performance
`
`
`SililerStar Exhibit 1016 - 18
`SilverStar Exhibit 1016 - 18
`
`

`

`4
`
`Sensors for Mobile Robots
`
`validation is invaluable in matching the capabilities and limitations of a particular
`sensor technology to the application at hand. Most manufacturers of established
`product
`lines provide excellent background information and experienced
`applications engineers to assist in this regard. but some of the more recently
`introduced devices are understandably a bit behind the power curve in terms of
`their documented performance results.
`in addition to the general theory of sensor
`operation, therefore, this book attempts to provide the reader with some important
`exposure to the practical experiences and insights of system developers involved
`in this rapidly evolving field.
`
`1.2 The Robots
`
`I consider m