`
` FOR MOBILE
`OBOTS.
`
`
`
`
`
`THEORY AND APPLICATION ©
`
`SilverStar Exhibit 1016
`SilverStar Exhibit 1016
`
`

`

`ISBN 1-56441-048-2
`
`SilverStar Exhibit 1016 - 2
`SilverStar Exhibit 1016 - 2
`
`

`

`Sensors
`for Mobile
`Robots
`
`Theory and Application
`
`H.R. Everett
`Naval Command, Control and
`Ocean Surveillance Center
`San Diego, California
`
`AKPeters, Ltd.
`Wellesley, Massachusetts
`
`SilverStar Exhibit 1016 - 3
`SilverStar Exhibit1016 - 3
`
`

`

`Editorial, Sales, and Customer Service Office
`
`AK Peters, Ltd.
`289 Linden Street
`Wellesley, MA 02181
`
`Copyright © 1995 by A K Peters, Ltd.
`
`All rights reserved. No part of the material protected by this copyright notice may
`be reproduced orutilized in any form, electronic or mechanical, including photo-
`copying, recording, or by any information storage andretrieval system, without
`written permission from the copyright owner.
`
`Library of Congress Cataloging-in-Publication Data
`
`Everett, H. R., 1949-
`Sensors for mobile robots : theory and application / H.R. Everett.
`p. cm.
`Includes bibliographical references and index.
`ISBN 1-56881-048-2
`1. Mobile robots. 2. Robots—Control systems.
`TJ211.415.E83 1995
`629.8 ' 92—dce20
`
`|. Title.
`
`95-17178
`cIP
`
`Many of the designations used by manufacturers and sellers to distinguish their products are
`claimed 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 efforl has been made to ensure accurate presentation of product names and
`specifics,
`
`Principle illustrator: Todd Ashley Everett
`
`Printed in the United States of America
`99 98 97 96 95 10987654321
`
`SilverStar Exhibit 1016 - 4
`SilverStar Exhibit 1016 - 4
`
`

`

`Table of Contents
`
`EEPIRESVO CERES msocxsaaracicesicnencecnacvasanctineneareeankeeecmeerenneeteremerteemerterereseseeXbti
`
`
`PREFACEvssssssesssseseinicio
`aaa ena
`iiss
`
`evseeeeXV
`
`NEKOWLEDGMENTS scscscssicessccscsccsostasctittioniesite SecelestSoetinicouteaiedecvebnebiestcaxvii
`
`ISINTRODCTION vopccccccccncorecercsiesumaiesiestencuceensenaarti cueamaasacee
`
`
`
`1.1 DESIGN CONSIDERATIONS........
`ae
`wa
`1.2 THE ROBOTS...
`
`
`1.2.1 WALTER(1965-1967)...
`cnc
`1.2.2 CRAWLERI (1966-1968)...
`eal
`
`
`1.2.3 CRAWLERI] (1968-197 1)....ssssssssseseescstrsesracseeteersneressennnsees
`
`1.2.4 ROBARTI (1980-1985) ..ue
`1.2.5 ROBARTII (1982-)..
`
`1.2.6 MODBOT(1990-) ..
`1.2.7 USMC TeleOperatedVehicle(1985- 1989)...
`
`1.2.8 MDARSInterior (1989-) .....0ccc..ccc e
`
`1.2.9 Surrogate Teleoperated Vehicle (1990-1993) ...
`wed
`1.2.10 ROBARTTI(1992-).....ccsescscsseceeceessesesse
`28
`1.2.11 MDARSExterior (11994-)..
`
`
`Lest canteen
`1.3 REFERENCES...
`
`2, DEAD RECKONING,.......
`
`
`2.1 ODOMETRY SENSORS.......
`
`2.1.1 Potentiometers....
`
`
`2.1.2 Synchros and Resolvers.
`er
`2.1.3 Optical Encoders ..
`2.2 DOPPLER AND INER’TIALNAVIGATION...
`
`2.2.1 Doppler Navigation...
`
`wae
`2.2.2 Inertial Navigation...
`2.3 TYPICAL MOBILITY CONFIGURATIONS
`
`2.3.1 Differential Steering..
`
`2.3.2 Ackerman Steering .
`
`2.3.3 Synchro Drive.....
`
`
`er
`2.3.4 Tricycle Drive..
`2.3.5 Omni-DirectionalDave...
`ae
`im
`
`2.4 INTERNAL POSITION ERRORCORRECTION,
`
`2:5 REFERENCES:sicctiscececsisissecsataciacnscath\sepincaniacaciaciel
`3. TACTILE AND PROXIMITY SENSING j....sssscscssssesssssssneesssrceeessnconsnsnnsnnensesssesseenensnesersegeesOF
`
`S21 TAGTSE: SENSORS.ss cciasnraciassccntanasiaatueson sssciteein fotceeav naantava iraa nla Foe ee a
`SF TC EG OPE acca geeneamenrens
`3.1.2 Tactile Bumpers vc.
`3.1.3 Distributed Surface Arrays...
`
`
`
`
`SilverStar Exhibit 1016 - 5
`SilverStar Exhibit 1016 - 5
`
`

`

`vi
`
`Sensors for Mobile Robots
`
`3.2 PROXIMITY SENSORS...
`
`
`3.2.1 Magnetic Proximity Sensors
`3.2.2 Inductive Proximity Sensors......
`
`3.2.3 Capacitive Proximity Sensors...
`
`3.2.4 Ultrasonic Proximity Sensors .
`
`3.2.5 Microwave Proximity Sensors...
`3.2.6 Optical eeSensors...
`
`3.3 REFERENCES...
`ee
`4, TRIANGULATION RANGING....essssseessserseesines tkesc asccantenehsacearnercactienaneeneayexatcnasanwes 103
`
`4.1 STEREO DISPARITY...
`
`4.1.1 JPL Stereo Vision...
`
`4.1.2 David Sarnoff Stereo Vision
`
`4.2 ACTIVE TRIANGULATION........
`
`4.2.1 Hamamatsu Rangefinder ChipSet.
`
`4.2.2 Draper Laboratory Rangefinder.........
`
`4.2.3 Quantic Ranging System.....
`
`4.3 ACTIVE STEREOSCOPIC..
`
`43.1 HERMIES..
`43.2 Danl-Aparturs3DRange Sensor...
`
`4.4 STRUCTURED LIGHT..
`
`4.4.1 TRC Strobed-Light Triangulation‘System...
`=
`4.5 KNOWN TARGETSIZE..
`we
`4.5.1 NAMCOLasernet™ Seanning Lasereects
`
`4.6 OPTICAL FLOW ..
`ii
`4.6.1 NIST Passive nies ae Collision ‘Avoidance.,
`
`4.6.2 David Sarnoff Passive Vision.....
`
`4.) REFERENCES5 jc0ccjjaiaseenicepiineinas
`
`
`Bi TIME OR FUMGEID isisssiccssssccecsnssaiseassetascubonssitianteyevesansaneassepnecevetaciesserouamrecavensbaqacsiosencenesesets 139
`
`5.1 ULTRASONIC TOF SYSTEMS...
`
`5.1.1 National Semiconductor's LM1812UltrasonicTransceiver.
`
`5.1.2 Massa Products Ultrasonic Ranging Module Subsystems...
`
`5.1.3 Polaroid Ultrasonic Ranging Modules.........0...004
`
`5.1.4 Cybermotion CA-2 Collision Avoidance System
`
`5.2 LASER-BASED TOF SYSTEMS......
`50
`
`
`5.2.1 Schwartz Electro-Optics Laser Rangefi
`«150
`5.2.2 RIEGL Laser Measurement Systems...
`158
`5.2.3 Odetics Fast Frame Rate 3-D Laser TihagingSystem.
`«161
`
`. 162
`3.2.4 RVSI LongERE iiiiatand Detectionmy Sse
`
`
`«165
`scngaaremmante
`§.3 REFERENCES..
`.
`
`6. PHASE-SHIFT MEASUREMENT AND FREQUENCY MODULATION.cscs 169
`
`6.1 PHASE-SHIFT MEASUREMENT.....
`
`6.1.1 ERIM 3-D Vision Systems..
`
`6.1.2 Perceptron LASAR........006
`
`6.1.3 Odetics Scanning Laser Imaging System...
`6.1.4 Sandia Scannerless Range Imager............
`
`6.1.5 ESP Optical Ranging System ......
`
`6.1.6 Acuity Research AccuRange 3000,
`
`
`SilverStar Exhibit 1016 -6
`SilverStar Exhibit 1016 - 6
`
`

`

`Table of Contents
`
`Vii
`
`6.1.7 TRC Light Direction and Ranging System...
`6.2 FREQUENCY MODULATION...
`6.2.1 VRSS Automotive Collision Avotlamee:Radar...
`
`6.2.2 VORADVehicle Detection and Driver Alert Sysian..
`6.2.3 Safety First Systems Vehicular Obstacle Detection andWarming System...
`
`6.2.4 Millitech Millimeter Wave Radar...
`6.3 REFERENCES. ..0...cccceccsccseseseetecesnenensaeeeeee
`
`
`
`
`7. OTHER RANGING TECHNIQUES...
`
`
`7.1 INTERFEROMETRY...
`—
`7.1.1 CLS Coordinate MeasuringSystem
`
`7.2 RANGE FROM FOCUS...
`tt
`
`7.2.1 Honeywell AnrafoousSyemene -
`
`7.2.2 Associates and Ferren Swept-Focus Ranging...
`7.2.3 JPL Range-from-Focus System........06
`
`7.3 RETURN SIGNAL INTENSITY ...........
`7.3.1 Programmable Near-Infrared Proximity Sensor......
`
`7.3.2 Australian National University Rangefinder.....
`7.3.3 MIT Near-Infrared Ranging System.......0...0..
`
`7.3.4 Honeywell Displaced-Sensor Ranging Unit.
`
`Ta REPRRENGRS ceninsknannnaes
`SA COUSTICAE ENERGY ssicccscsivivcinesntacssaveceisassnesonrsscevsinsetsebboaseneesieneestossescaveltaeivonitenyseoienBOD
`
`
` 8.1 APPLICATIONS ..
`8.2 PERFORMANCE FACTORS..
`225
`
`8.2.1 Atmospheric Atlenuation.........cccccsececoe
`225
`
`B22 Tarstet REnestiv ty cccsccccssecsaisavsnosvstiavisocavans
`227
`
`8.2.3 Air Turbulence...
`232
`
`
`8.2.4 Temperature...
`233
`
`8.2.5 Beam Geometry...
`234
`DUBAD INGING au cncasistaintieaeiainahitins
`
`8.2.7 System-Specific Anomialies..
`8.3 CHOOSING AN OPERATING FREQUENCY
`
`8.4 SENSOR SELECTION CASE STUDY..
`
`HOS
`8.5 REFERENCES...
`
`9. ELECTROMAGNETIC ENERGY........
`
`
`
`5
`
`9.1 OPTICAL ENERG¥oiviveticsteceetyesssesceees
`9.1.1 Electro-Optical Sources...
`9.1.2 Performance Factors...
`9.1.3 Choosing an OperatingWavelength.
`
`9.2 MICROWAVE RADAR..
`aoe
`
`9.2.1 Applications ...
`
`9.2.2 Performance Factors......
`
`9.3 MILLIMETER-WAVERADAR..
`
`9.3.1 Applications...
`9.3.2 Performance Factors...
`oats
`9.3.3 ChoosinganOperatingFrequency.
`
`Fr
`
`9.4 REFERENCES...
`
`aang
`
`
`SilverStar Exhibit 1016 - 7
`SilverStar Exhibit 1016 - 7
`
`

`

`Vill
`
`Sensors for Mobile Robots
`
`10. COLLISION AVOIDANCE,....cccccnsessecsssreeessFead.veesnbaesentensunanpoatecaundeealplccssiorteteeieaenenaarEee
`
`
`10.1 NAVIGATIONAL CONTROL STRATEGIEG..........
`279
`10.1.1 Reactive: Controliiaiecasic.cvivcnieisieccitiees
`280
`
`290
`[0.1.2 Representational World Modeling.
`
`10.1.3 Combined-Approach ccicixi...21.dscecateccteecase
`on 295
`10.2 EXTERIOR APPLICATION CONSIDERATIONS...
`on 299
`
`10.3 NAVIGATIONAL RE-REFERENCING.............
`anu
`
`1 dO2
`10:4: REFERENCES) esicinicssaccsncn
`
`
`11. GUIDEPATH FOLLOWING ocscssscissesesssesscenssnesssacbieciciesiesenelNESSUSBiceeasiestSOD
`
`EST WORE) GUIDED issca. sosaveascaicaccicstectenisnnntaaniansctansstessonacien
`11.2 OPTICAL STRIPE.......
`
`
`11.2.1 ModBot Optical Stripe Tracker............
`
`11.2.2 U/V Stimulated Emission.........0.:0.00-
`
`11.3 MAGNETIC TAPE. ............csscceeceseese
`
`11.3.1 Macome Magnetic Stripe Follower...
`11.3.2 Apogee Magnetic Stripe Follower...
`
`11.3.3 3M/Honey well Magnetic Lateral Guidance System
`11.4 HEAT AND ODOR SENSING..,....cssesccsses
`11.5 INTERMITTENT-PATH NAVIGATION...
`
`
`11.5.1 MDARSInterior Hybrid Navigation........
`322
`ie
`11.5.2 Free Ranging On Grid.....
`
`
`116 REPRERENCESeossccrssreesaieanesconesscocansaianatatcetsennsstcoccaanteneansarvasanvatucusntesncauericesseasxstaeranniontbal325
`
`
`306
`
`
`
`TZ. MAGNETIC COMPASSIS wisesisivesiensssrenitinniessciysenniietnahnsenanantyieeiestaniansabiiuetbdbentenetshabaies 327
`
`12.1 MECHANICAL MAGNETIC COMPASSES.....cesscessccsccscsesresessecensnsteenttsadbeatsnadeneesneeneereaens 2B
`
`12.1.1 Dinsmore Starguide Magnetic Compass...
`x
`dd
`
`12.2 FLUXGATE COMPASSES............
`330
`
`12.2.1 Zemeo Fluxgate Compasses
`337
`12.2.2 Watson Gyro Compass...
`340
`
`12.2.3 KVH Fluxgate Compasses
`4]
`12.2.4 Applied Physics Systems Miniature Orientation Sensor:
`1.343
`
`12.3 MAGNETOINDUCTIVE MAGNETOMETERS..
`garsub.feaunige nay eaciunamraneaLanganayens sary 344
`12.3.1 Precision Navigation TCM Meencisinductive———e URN Mees 345
`
`12.4 HALL-EFFECT COMPASSES.........00:
`347
`
`349
`12.5 MAGNETORESISTIVE COMPASSES.
`
`350
`12.5.1 Philips AMR Compass............
`351
`
`12.5.2 Space Electronics AMR Compass...
`
`12.5.3 Honeywell HMRSeries Smart peekMagncicenctet
`oan
`
`12.6 MAGNETOELASTIC COMPASSES..
`ea ra aera
`TRIER 353
`EET REFERENCES sccsissrcctos eaecennnicornanranane 3357
`
`TR. GY ROSCOPES cpesiscosivccosssssevcsncsuenusinissiecasansrevssrstotocsoveatenenisteniteesssersctertisstsicneaceivieaebieonsieenSOL
`
`
`
`13.1 MECHANICAL GYROSCOPES...........
`si seincaean aaisnTistaaaasi 362
`13.1.1 Space-Stable Gyroscopes.
`
`13.1.2 GyfOCOMPASSES.......cccsesececcesieessssens
`sheer
`sscdCasoavbs ap acme taka
`13.1.3 Rate GYTr0S...........cscccsseeressssssseseneseeseenereesearssinetsanenentnesetsesertenesatsseetnesarsesersenesserreatenss JOD
`
`
`SilverStar Exhibit 1016 - 8
`SilverStar Exhibit 1016 - 8
`
`

`

`Table of Contents
`
`Ix
`
`13.2 OPTICAL GYROSCOPES .....0000.....
`
`13.2.1 Active Ring-Laser Gyros......
`
`13.2.2 Passive Ring Resonator GyroS.............ccececeeseseeee
`
`13.2.3 Open-Loop Interferometric Fiber-Optic Gyros.
`13.2.4 Closed-Loop Interferometric Fiber-Optic Gyros...
`
`13.2.5 Resonant mn ee
`
`13.3 REFERENCES...
`sa
`e
` 14. RF POSITION-LOCATION SYSTEMS .........+0100
`
`Shine ION
`
`astancaa eRiea
`
`14.1 GROUND-BASEDRF SYSTEMS...
`14.1.1 Loran...
`
`14.1.2 Kaman SciencesRadioFrequency Navigation Grid.........
`[4.1.3 Precision Technology Tracking andterete _—
`14,1.4 Motorola Mini-Ranger Falcon...
`Jaaviiyontnneitoaniaats
`
`
`14.1.5 Harris Infogeometric System...
`14.2 SATELLITE-BASED SYSTEMS....
`
`14.2.1 Transit Satellite Navigation System
`14.2.2 Navstar Globalplaitieditie areas
`
`
`14.3 REFERENCES...
`Pretec ere
`15. ULTRASONIC AND OPTICAL POSITION-LOCATION SYSTEMS,Q.......:-s:ssseseeee+00423
`
`
`15.1 ULTRASONIC POSITION-LOCATION SYSTEMS |. iisccsicisescsicssastlcenesnidencietinciecntiod 423
`
`
`15.1.1 Ultrasonic Transponder Trilateration .........0cc0sce
`vee hd
`
`431
`15.1.2 Ultrasonic Signature Matching...
`433
`15.2 OPTICAL POSITION-LOCATION SYSTEMS.
`
`433
`15.2.1 CRAWLER| Homing Beacon..........
`
`434
`15.2.2 ROBARTII Recharging Beacon..
`
`436
`15.2.3 Cybermotion Docking Beacon..
`
`14.2.4 Hilare..
`oe
`
`15:25 NAMCOLaseuieta StunningLiaeee Sensor...
`15.2.6 Caterpillar Self-Guided Vehicle...
`:
`
`15.2.7 TRC Beacon Navigation Syitetn...aaiviqaatters
`15.2.8 Intelligent Solutions EZNav Position Sensor.....
`15.2.9 Imperial College Beacon Navigation System.....
`
`15.2.10 MTI Research CONAC..............
`
`15,2.11 MDARSLateral-Post Sensor...
`
`15.3 REFERENCEG..........
`
`
`16. WALL, DOORWAY, AND CEILING REFERENCING. ......ccccccssscsscssceenteteessestseseseaeen455
`
`Seas
`16.1 WALL REFERENCING.....
`
`16.1.1 Tactile Wall Referencing...
`16.1.2 Non-Contact Wall Referencing .
`
`16.1.3 Wall Pollowing.......csecee
`16.2 DOORWAY TRANSIT REFERENCING
`
`16.3 CEILING REFERENCING..
`a
`16.3.1 Polarized Optical Heading Referenceeee
`
`16.3.2 Georgia Tech Ceiling Referencing System....
`
`16.3.3 TRC HelpMate Ceiling Referencing System...
`
`16.3.4 MDARS Overhead-Beamsaeae
`
`16.4 REFERENCES...
`"i
`
`
`SilverStar Exhibit 1016 - 9
`SilverStar Exhibit 1016 - 9
`
`

`

`x
`
`Sensors for Mobile Robots
`
` 17, APPLICATION-SPECIFIC MISSION SENSORS...
`
`17,1 THE SECURITY APPLICATION..
`
`
`17.1.1 Acoustical Detection .
`17.1.2 Vibration Sensors..........
`
`17.1.3 Ultrasonic Presence Sensors.
`
`17.1.4 Optical Motion Detection...........006
`
`17.1.5 Passive Infrared Motion Detection.
`17.1.6 Microwave Motion Detection................
`
`17.1.7 Video Motion Detection .........0....
`
`17.1.8 Intrusion Detection on the Move,
`
`17.1.9 Verification and Assessment..........
`
`17.2 AUTOMATED INVENTORY ASSESSMENT...
`
`17.2.1 MDARSProduct Assessment System ....
`
`17.3 REFERENCES.........0.:00004
`
`
`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 right
`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 offiction.
`
`Sensors are what makesit 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 computervision for robots.
`There is exactly one book devoted to non-visual sensors. This one.
`Wetend 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 (hmmm...) begins to emerge.
`Insects have two eyes, each with at most
`perhaps 10,000 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 andalter the direction
`in which they are scuttling 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 al
`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 ownlittle libraries of sensors in our heads. Now Bart Everett has written
`downall he had in his own private library and more. Bart's robots have always
`stood out as those with the most sensors, because interactive sensing has always
`been a priority for Bart. Now he is sharing his accumulated wisdom with us, and
`robotdom 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 whatits limitations will
`be.
`
`So let’s go build some new robots!
`
`Rodney A. Brooks
`MIT AI Lab
`Cambridge, MA
`
`
`SilverStar Exhibit 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 ofinformation on al] 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 underwater 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 knownto exist; one of these methods alone (triangulation)
`can be implementedin 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, | 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 |, 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, phase-shift
`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
`image-based systems, as the subject of machine vision quite obviously could be
`the focus of a book all in itself. And since a numberofdistinguished 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
`becomesratherdetailed.
`I have very much enjoyed the preparation of this manuscript, both in terms of
`what I learned in the process and the new contacts | 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 awarenessofavailable supporting technologies.
`
`H.R. 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 meto electronics at
`an early age.
`
`My high school geometry teacher, Mrs. Nell Doar, for providing discipline,
`inspiration, and the mathematical foundation upon which I wasto build.
`
`Professor Robert Newton, my thesis advisor at the Naval Postgraduate School,
`who made it possible for me to pursue a rather unorthodox 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 school.
`
`Dr. Anita Flynn of MIT forall the late nights and weekends we spent hacking
`code and building our own sensors in my basement in Virginia.
`
`Gary Gilbreath of the Naval Command Control and Ocean Surveillance Center
`for transforming ROBARTII 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 manuscriptin 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 Hower, Robin Laird, Richard Langley, Richard
`Lao, Larry Mathies, and Hoa Nguyen.
`
`In addition, portions of the material presented in Chapters 4 through 7 were
`previously published in Sensers and later Robotics and Autonomous Systems
`magazines, and updated in this book with their kind permissions.
`
`SilverStar Exhibit 1016 - 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. Additiona) 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:
`1) 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 havetransitioned from 6502- and Z80-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
`full-duplex Ethernet-compatible high-speed datalinks with ranges of several
`miles.
`
`
`SilverStar Exhibit 1016-16 |
`SilverStar Exhibit 1016 - 16
`
`

`

`2
`
`Sensors for Mobile Robots
`
`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-
`off differential or contrast. Unfortunately, such optimized configuration controlis
`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 real-world 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 of 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 oninterior systems,
`On the positive side, worldwide interest in a rapidly expanding field known as
`intelligent vehicle highway systems (IVHS) 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 low-cost 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
`technology-transfer process.
`
`1.1 Design Considerations
`
`The problems confronting most mobile robotic developmentefforts 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:
`
`e
`
`Field of view — Should be wide enough with sufficient depth of field to
`suit the application.
`e Range capability — The minimum range of detection, as well as the
`maximum effective range, must be appropriate for the intended use of the
`sensor.
`
`e
`
`* Accuracy and resolution — Both must be in keeping with the needs ofthe
`given task.
`« 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.
`e Real-time operation — The update frequency must provide rapid, real-
`time data at a rate commensurate with the platform’s speed of advance
`(and take into accountthe velocity of other approaching vehicles).
`e Concise, easy to interpret data — The output format should berealistic
`from 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,
`e 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 ofall targets, as well as
`to increase the confidence level of the output.
`Simplicity — 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.
`
`«
`
`*
`
`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
`
`
`SilverStar 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 prov