`
`NAVAL POSTGRADUATE SCHOOL
`Monterey, California
`
`DISSERTATION
`
`INERTIAL AND MAGNETIC TRACKING OF LIMB
`SEGMENT ORIENTATION FOR INSERTING HUMANS
`INTO SYNTHETIC ENVIRONMENTS
`
`by
`
`Eric Robert Bachmann
`
`December 2000
`
`Dissertation Supervisor:
`Approved for public release; distribution is unlimited.
`
`Michael J. Zyda
`
`DTIC QUt .. LJ.:rY I!Ja?ElCTED 1
`
`200,022, 078
`
`Google 1047
`
`
`
`REPORT DOCUMENTATION PAGE
`Form Approved
`0MB No. 0704-0188
`Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instruction,
`searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send
`comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to
`Washington headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA
`22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188) Washington DC 20503.
`
`1. AGENCY USE ONLY (Leave blank)
`
`2. REPORT DATE
`December 2000
`
`3. REPORT TYPE AND DATES COVERED
`Ph.D. Dissertation
`
`4. TITLE AND SUBTITLE
`Inertial and Magnetic Tracking of Limb Segment Orientation for Inserting Humans into
`Svnthetic Environments
`6. AUTHOR(S)
`Bachmann, Eric R.
`
`7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
`Naval Postgraduate School
`Monterey, CA 93943-5000
`
`5. FUNDING NUMBERS
`ARO Proposal No.
`40410-MA
`
`N0003900WRDR053
`
`8. PERFORMING
`ORGANIZATION REPORT
`NUMBER
`
`9. SPONSORING/ MONITORING AGENCY NAME(S) AND ADDRESS(ES}
`U.S. Army Research Office (ARO) Research Triangle Park, NC 27709-2211
`U.S. Navy Modeling and Simulation Office (N6M) Washington, DC 20350-2000
`11. SUPPLEMENTARY NOTES
`The views expressed in this thesis are those of the author and do not reflect the official policy or position of the Department of
`Defense or the U.S. Government.
`12a. DISTRIBUTION/ AVAILABILITY STATEMENT
`
`10. SPONSORING/
`MONITORING
`AGENCY REPORT NUMBER
`
`12b. DISTRIBUTION CODE
`
`Approved for public release; distribution is unlimited.
`13. ABSTRACT (maximum 200 words)
`Current motion tracking technologies fail to provide accurate wide area tracking of multiple users without interference and
`occlusion problems. This research proposes to overcome current limitations using nine-axis magnetic/angular rate/gravity (MARG)
`sensors combined with a quaternion-based complementary filter algorithm capable of continuously correcting for drift and
`following angular motion through all orientations without singularities.
`Primarily, this research involves the development of a prototype tracking system to demonstrate the feasibility of MARG
`sensor body motion tracking. Mathematical analysis and computer simulation are used to validate the correctness of the
`complementary filler algorithm. The implemented human body model utilizes the world-coordinate reference frame orientation data
`provided in quaternion form by the complementary filter and orients each limb segment independently. Calibration of the model and
`the inertial sensors is accomplished using simple but effective algorithms. Physical experiments demonstrate the utility of the
`proposed system by tracking of human limbs in real-time using multiple MARG sensors.
`The system is "sourceless·· and does not suffer from range restrictions and interference problems. This new technology
`overcomes the limitations of motion tracking technologies currently in use. It has the potential to provide wide area tracking of
`multiple users in vinual en\'lronment and augmented realitv annlications.
`14. SUBJECTTERMS
`Micromachined Sensors. Complemenwy Filtering, Quaternions, Motion Tracking, Networked Virtual
`Environments, Quatemion/V ector Pairs, Human Modeling, MARG Sensors, Inertial Sensors, Magnetic Sensors
`
`15. NUMBER OF
`PAGES
`195
`
`17. SECURITY CLASSIFICATION OF
`REPORT
`Unclassified
`NSN 7540-01-280-5500
`
`18. SECURITY CLASSIFICATION OF
`THIS PAGE
`Unclassified
`
`19. SECURITY CLASSIFI- CATION
`OF ABSTRACT
`Unclassified
`
`16. PRICE CODE
`20. LIMITATION
`OF ABSTRACT
`UL
`Standard Form 298 (Rev. 2-89)
`Prescribed by ANSI Std. 239-18
`
`
`
`ii
`
`
`
`Approved for public release; distribution is unlimited
`
`INERTIAL AND MAGNETIC ANGLE TRACKING OF LIMB SEGMENTS
`FOR INSERTING HUMANS INTO SYNTHETIC ENVIRONMENTS
`
`by
`
`Eric Robert Bachmann
`B.A., University of Cincinnati, 1983
`M.S., Naval Postgraduate School, 1995
`
`Submitted in partial fulfillment of the
`requirements for the degree of
`
`DOCTOR OF PIDLOSOPHY IN COMPUTER SCIENCE
`
`from the
`
`NAVAL POSTGRADUATE SCHOOL
`December 2000
`
`Author:
`
`Approved by:
`
`Michael J. Zyd
`
`W4{p,\[~
`
`Robert B. McGhee
`Professor of Computer Science
`
`Man-Tak Shi g, Associate
`Professor of Computer Science
`
`Approved by:
`
`Approved by:
`
`iii
`
`Computer Science
`
`Xiaoping Yun,
`Professor of Electrical Engineering
`
`C\e ◊ • C6-:::'\~ L c:::::)0/
`Don~£, Assistant F
`
`Professor of Applied Science
`
`
`
`
`
`IV
`
`
`
`ABSTRACT
`
`Current motion tracking technologies fail to provide accurate wide area tracking of
`
`multiple users without interference and occlusion problems. These limitations make
`
`difficult the construction of a practical and intuitive interface, which allows humans to be
`
`inserted into networked virtual environments in a fully immersive manner. Advances in the
`
`field of miniature sensors make possible inertial/magnetic tracking of human body limb
`
`segment orientation without the limitations of current systems. Due to implementation
`
`challenges, inertial/magnetic sensors have not previously been used successfully for full
`
`body motion capture. This research proposes to overcome these challenges using multi-axis
`
`sensors combined with a quaternion-based complementary filter algorithm capable of
`
`continuously correcting for drift and following motion through all orientations without
`
`singularities.
`
`Primarily, this research involves the development of a prototype tracking system to
`
`demonstrate the feasibility of hybrid RF/magnetic/inertial motion tracking. Construction of
`
`inertial/magnetic (MARG) sensors is completed using off-the-shelf components.
`
`Mathematical analysis and computer simulation are used to validate the correctness of the
`
`complementary filter algorithm. The implemented human body model utilizes the world(cid:173)
`
`coordinate reference frame orientation data provided in quaternion form by the
`
`complementary filter and orients each limb segment independently. Calibration of the
`
`model and the inertial sensors is accomplished using simple but effective algorithms.
`
`Physical experiments demonstrate the utility of the proposed system. These experiments
`
`involve the tracking of human limbs in real-time using multiple inertial sensors.
`
`The motion tracking system produced has an accuracy which is comparable and a
`
`latency which is superior to active electro-magnetic sensors. The system is "sourceless"
`
`and does not suffer from range restrictions and interference problems. With additional
`
`MARG sensor packages, the architecture produced will easily scale to full body tracking.
`
`This new technology overcomes the limitations of motion tracking technologies currently
`
`V
`
`
`
`in use. It will provide wide area tracking of multiple users in virtual environment and
`
`augmented reality applications.
`
`vi
`
`
`
`TABLE OF CONTENTS
`
`I.
`
`II.
`
`1.
`2.
`3.
`
`INTRODUCTION .................................................................................................... 1
`A. MOTIVATION ................................................................................................ 1
`GOALS ............................................................................................................ 5
`B.
`1.
`Problem to be Solved ............................................................................. 5
`2. What is Fundamentally New .................................................................. 5
`3.
`Contribution of this Research ................................................................ 6
`l\IBTHOD ........................................................................................................ 6
`C.
`DISSERTATION ORGANIZATION ............................................................. 7
`D.
`SURVEY OF TRACKING TECHNOLOGIES ....................................................... 9
`A.
`INTRODUCTION ........................................................................................... 9
`B. MOTION TRACKING TECHNOLOGIES .................................................... 9
`a.Framework for Suitability .................................................................. 9
`b.Performance Requirements .............................................................. 11
`Mechanical Trackers ............................................................................ 13
`Magnetic Trackers ................................................................................ 15
`Optical Sensing .................................................................................... 19
`a.Pattern Recognition Systems .......................................... : ................ 20
`b.lmage Based Systems ...................................................................... 21
`c.Structured Light and Laser Systems ................................................ 24
`Acoustic Trackers ................................................................................ 25
`4.
`Inertial and Magnetic Tracking ............................................................ 25
`5.
`RF Positioning ...................................................................................... 28
`6.
`Hybrid Tracking Systems ..................................................................... 29
`7.
`8.
`Other Technologies .............................................................................. 31
`SUMMARY ................................................................................................... 32
`C.
`ID. REPRESENTATION OF HUMAN BODY MOTION AND MODELING ......... 33
`A.
`INTRODUCTION ......................................................................................... 33
`RIGID BODY ORIENTATION REPRESENTATION ................................ 33
`B.
`1.
`Euler Angles ......................................................................................... 34
`a.Euler Angle Rotation ....................................................................... 35
`b.Transforming Body Rates To Euler Rates ....................................... 36
`c.Euler Angle Singularities ................................................................. 38
`Quaternions .......................................................................................... 39
`a.Quaternion Operations ..................................................................... 40
`b.Quaternion Forms ............................................................................ 41
`c.Quaternion Transformation Between Coordinate Frames .............. .42
`d.Unit Quaternions In Positive Real Form .......................................... 44
`e.Transforming Angular Rates To A Quaternion Rate ....................... 44
`f.Representing Orientations Without Singularities ............................ .46
`C. MODELS FOR HUMAN BODY TRACKING ........................................... .46
`Kinematic Models Based On Homogenous Transformation Matrices 47
`1.
`
`2.
`
`vii
`
`
`
`F.
`
`Forward and Inverse Kinematics ......................................................... 48
`2.
`Kinematic Models of the Human Body based on Joint Angles .......... .49
`3.
`Orientation Only Tracking ................................................................... 51
`4.
`Kinematic Models based on Quaternion/Vector Pairs ........................ 51
`5.
`SUMMARY AND CONCLUSIONS ............................................................ 55
`D.
`IV. REVIEW OF FILTER THEORY AND DESIGN ................................................ 57
`A.
`INTRODUCTION ......................................................................................... 57
`B. MINIATURE INERTIAL SENSORS ........................................................... 58
`C.
`RANDOM PROCESSES .............................................................................. 60
`D.
`LEAST SQUARES FII..'TERING .................................................................. 62
`E. WIENER FIL'TERING .................................................................................. 64
`1.
`Continuous Weiner Filters ................................................................... 65
`2.
`Discrete Weiner Filters ........................................................................ 67
`KALMAN FILTERING ................................................................................ 68
`1.
`Discrete Kalman Filters ....................................................................... 69
`2.
`Extended and Linearized Kalman Filters ............................................. 71
`G. COMPIBMENTARY FII..TERING .............................................................. 73
`1.
`Crossover Frequency ............................................................................ 76
`SUMMARY AND CONCLUSIONS ............................................................ 77
`H.
`V. A QUA'TERNION ATTITUDE FILTER .............................................................. 81
`A.
`INTRODUCTION ......................................................................................... 81
`B. A QUA'TERNION ATTITUDE FII..TER ...................................................... 81
`1.
`Parameter Optimization ....................................................................... 83
`2.
`Analysis ................................................................................................ 86
`a.Noise Response ................................................................................ 87
`b.Response to Initial Condition Errors ............................................... 87
`c.Choosing the Feedback gain value ................................................... 89
`Reduced Order Filter ............................................................................ 91
`a.Orthogonal Quaternion Theorem ..................................................... 92
`Differential Weighting of Sensor Data ................................................ 95
`4.
`Reduced Rate Drift Correction ............................................................ 96
`5.
`FII..TER SIMULATION ................................................................................ 97
`C.
`SUMMARY ................................................................................................... 98
`D.
`IMPLEMENTATION OF INERTIAL AND MAGNETIC TRACKING OF
`ffiJMAN LIMB SEGMENTS ............................................................................. 101
`A.
`INTRODUCTION ....................................................................................... 101
`B.
`PROTOTYPE MARG SENSORS ............................................................... 102
`1.
`Sensor Components ............................................................................ 104
`a.Crossbow CXL04M3 Triaxial Accelerometer ......................... : ..... 104
`b.Tokin CG-16D Series Rate Gyros ................................................. 104
`c.Honeywell HM:C2003 3-Axis Magnetometer ................................ 105
`2. Magnetometer Set/Reset .................................................................... 106
`3.
`Analog to Digital Conversion ............................................................ 107
`
`3.
`
`VI.
`
`viii
`
`
`
`C.
`
`Data Processing .................................................................................. 107
`4.
`SYSTEM SOFTWARE ............................................................................... 108
`Quaternion Filter ................................................................................ 113
`1.
`Sensor Calibration .............................................................................. 115
`2.
`Quaternion Human Body Model ........................................................ 121
`3.
`a.Setting Model Position and Posture ............................................... 123
`b.Body Model Calibration ................................................................ 126
`D. S~RY ................................................................................................. 130
`VII. EXPERII\IBNTAL RESlJI..TS ...................................................................... : ...... 131
`INTRODUCTION ....................................................................................... 131
`A.
`STATIC STABII.ITY .................................................................................. 131
`B.
`STA TIC CONVER.GENCE ........................................................................ 135
`C.
`D. DYNAMIC RESPONSE AND ACCURACY ............................................ 138
`QUALITATIVE TESTING ......................................................................... 138
`E.
`1. Weighted Least Squares ..................................................................... 138
`Posture Estimation ............................................................................. 139
`2.
`Reduced Rate Drift Correction .......................................................... 141
`3.
`INTERSENSE INERTIACUBE .................................................................. 143
`F.
`SUMMARY ................................................................................................. 145
`G.
`VIII. SUMMARY AND CONCLUSIONS ................................................................... 147
`INTRODUCTION ....................................................................................... 147
`A.
`B. MARG SENSORS ....................................................................................... 147
`HUMAN BODY MODELING .................................................................... 150
`C.
`INTERGRATION OF INERTIAL AND RF TECHNOLOGIES ............... 152
`D.
`E. WIRELESS COMMUNICATIONS ............................................................ 153
`FILTERING ................................................................................................. 154
`F.
`G. A PROTOTYPE INERTIAL TRACKING BODY SUIT ........................... 155
`POSTURE DATA IN A NETWORKED SYNTHETIC
`H.
`ENVIRONMENT ........................................................................................ 156
`CONCLUSIONS ......................................................................................... 157
`I.
`APPENDIX A.DERIVATION OF GAUSS-NEWTON ITERATION EQUATIONS 159
`APPENDIXB.DERIVATIONOFTHEXMATRIX ................................................. 161
`APPENDIX C. VIDEO DEMONSTRATION ............................................................. 165
`UST OF REFERENCES ................................................................................................ 167
`INITIAL DISTRIBUTION LIST ................................................................................... 175
`
`ix
`
`
`
`X
`
`
`
`LIST OF FIGURES
`
`Figure 1 : Exoskeleton tracking of the upper body .......................................................... 14
`Figure 2 : Electromagnetic Orientation Only Tracking of the Human Body
`From [Ref. 78.] ....................................... : ...................................................... 18
`Figure 3 : Frame Assignment Under MDH After [Ref. 17 .] .......................................... .48
`Figure 4 : Inertial Motion Tracking of the Right Fore and Upper Arm with Two Inertial
`Sensors and a Quaternion Attitude Filter From [Ref. 88.] ............................ 52
`Figure 5 : Human Model Designed For Quaternion Input ............................................... 53
`Figure 6 : Block Diagrams of Linear Systems ................................................................. 62
`Figure 7: Kalman Filter Loop After [Ref. 14.] ............................................................... 71
`Figure 8: Complementary Filter Block Diagram ............................................................ 74
`Figure 9 : Transform Domain Block Diagram Of Roll Angle Estimation Filter ............. 75
`Figure 10 : Quaternion-Based Attitude Filter From [Ref. 8.] .......................................... 82
`Figure 11 : Signal Flow Graph for Linearized System After [Ref. 54.] .......................... 86
`Figure 12: Simplified SFG For Static Testing With Zero Noise After [Ref. 55.] .......... 87
`Figure 13 : Transform Domain SFG For After [Ref. 55.] ............................................... 88
`Figure 14: Block Diagram Of Time Domain Linearized Quaternion Attitude Filter ..... 89
`Figure 15: Simulated Nonlinear Filter Response,10 Degree Offset, a=0.1, Dt=O.l From
`[Ref. 6.] ..................................................................................... · ..................... 98
`Figure 16 : Prototype Inertial and Magnetic Body Tracking System ............................ 102
`Figure 17 : Prototype MARG Sensor From [Ref. 61.] .................................................. 103
`Figure 18: MARG Sensor Magnetometer Set/Reset Circuit Schematic From [Ref. 61.] ...
`108
`Figure 19: Body Tracking Software Simplified Class Diagram ................................... 110
`Figure 20 : Class Instance Data Flow Diagram ............................................................. 112
`Figure 21 : Orientation Estimation Flow Chart ............................................................. 114
`Figure 22 : Dialog For Manually Setting Filter Parameters and Sensor Data Null Voltages
`and Scale Factors ......................................................................................... 115
`Figure 23: Rotating Sensor 90 Degrees About Positive x-axis For Rate Calibration ... 118
`Figure 24: Console Display Of Sensor Calibration Results ......................................... 121
`Figure 25 : Wireframe Rendering Of The Quaternion-Based Human Model ............... 122
`Figure 26 : Human Model Settings Dialog .................................................................... 123
`Figure 27 : Calculation Of Limb Segment Positions ..................................................... 125
`Figure 28 : The setPosture Method Of the CHumanModel Class ................................. 126
`Figure 29 : The renderFigure Method Of the CHumanModel Class ............................. 127
`Figure 30: Body Model Calibration Reference Position .............................................. 129
`Figure 31 : One Hour Static Test Of Orientation Estimate Stability, k = 1.0, = 1.0 ...... 132
`Figure 32 : 15 Minute Static Test Of Orientation Estimate Stability,
`No Magnetometer Input, k = 1.0 ................................................................. 133
`Figure 33 : 15 Minute Static Test Of Orientation Estimate Stability,
`No Accelerometer Input, k = 1.0 ................................................................. 133
`Figure 34 : 60 Minute Static Test Of Orientation Estimate Stability.
`
`xi
`
`
`
`No Rate Sensor Input, k = 1.0 ..................................................................... 134
`Figure 35 : Error Convergence Following 30 Degree Transient Error, k = 1.0 ............. 136
`Figure 36: Error Convergence Following 30 Degree Transient Error, k = 4.0 ............. 136
`Figure 37: Error Convergence Following 30 Degree Transient Error, k = 16.0 ........... 137
`Figure 38 : Error Convergence Following 30 Degree Transient Error, k = 32.0 ........... 137
`Figure 39: 10 Degree Roll Excursions At 10 deg/sec From [Ref. 6.] ........................... 139
`Figure 40 : rms Change In Orientation Estimate During Exposure Magnetic Source, Mag-
`netometer Weighting Factor: 1.0, k = 4.0 .................................................... 140
`Figure 41 : rms Change In Orientation Estimate During Exposure Magnetic Source, Mag-
`netometer Weighting Factor: 0.5, k = 4.0 .................................................... 140
`Figure 42 : rms Change In Orientation Estimate During Exposure Magnetic Source, Mag-
`netometer Weighting Factor: 0.25, k = 4.0 .................................................. 141
`Figure 43 : Inertial Tracking Of the Left Arm Using Three MARG Sensors ................ 142
`Figure 44 : Closed Kinematic Chain Posture Using Three MARG Sensors ................. 142
`Figure 45 : Inertial Tracking Of the Left Leg Using Three MARG Sensors ................. 143
`Figure 46 : Intersense InertiaCube ................................................................................. 145
`Figure 47 : MARG Rate Sensor Bias Compensation Circuit Schematic
`From [Ref. 61.] ............................................................................................ 149
`
`xii
`
`
`
`LIST OF TABLES
`
`Table 1: CXL04M3 Triaxial Accelerometer Specifications After [Ref. 18.] ................. 104
`Table 2: CG-16D Ceramic Rate Gyro Specifications After [Ref. 84.] ........................... 105
`Table 3: Honeywell HMC2003 Three-Axis Magnetic Sensor Hybrid
`Specifications After [Ref. 39.] .......................................................................... 106
`
`xiii
`
`
`
`
`
`
`
`XiV
`xiv
`
`
`
`ACKNOWLEDGEMENTS
`
`To achieve any worthwhile goal, all of us must stand on the shoulders of those who
`
`have come before and learn from our gifted contemporaries. Otherwise, we would all still
`
`be wearing animal skins and wondering what shape the wheel should have. Completion of
`
`this research was only possible due to the talent and patience of many special individuals.
`
`This dissertation serves as documentation of their contributions.
`
`Most of all I would like to express my gratitude to Dr. Robert McGhee. From
`
`beginning to end, his patience as well as sheer genius made possible any progress that was
`
`made. His intellectual contributions to this work as well as myself are tremendous.
`
`However, it is for his role as my supporter and his efforts to boost my confidence when it
`
`waned that I am most thankful. It is a measure of my respect for Dr. McGhee, that even
`
`after all the time we have spent together, that I am still unable to call him by his first name.
`
`During the period under which this work was completed, the advice and guidance of
`
`Michael Zyda did much to help me to limit the scope of the work undertaken and keep me
`
`focused on the 'real goal.' The real goal being of course to finish the dissertation and
`
`graduate. There were many extended periods during which I felt little or no progress was
`
`being made. I often waited for Mike to ask what the heck I was up to, but he was patient
`
`and simply waited for the results which he seemed to know would come even if I was not
`
`so sure.
`
`This is a dissertation for a doctorate in Computer Science. However, it involved the
`
`study of se\'eral topics that are normally associated with Electrical Engineering. I am not
`
`an electrical engineer. but through many hours, the patience and teaching skills of Dr.
`
`Xiaoping Yun brought me as close as I will ever be. Even though Xiaoping is extremely
`
`brilliant. he was still always able to come down to my level to successfully teach me the
`
`complexities of random processes, the frequency domain and optimal filtering. The
`
`difficulty of this acheivement should not be underestimated.
`
`xv
`
`
`
`In the beginning there was a qualifying exam ... It was full of automata and algorithms.
`
`Mantak Shing generously gave me his time and lent me his knowledge to help me to
`
`successfully prepare for the exam. The mind of Don Brotzman is truly amazing. I look at
`
`something and I am only able to see it one way. Don looks at the same thing and is able to
`
`see it from an infinite number of viewpoints. I am very grateful for his multi-dimensional
`
`advice and guidance.
`
`I have had the pleasure of advising numerous students who have contributed to this
`
`research. Among these, the contributions of Ildeniz Duman and Umit Usta were
`
`particularly significant. Funding was provided by the U.S. Army Research Office (ARO),
`
`the U.S. Navy Modeling and Simulation Office (N6M), and the Naval Postgraduate School.
`
`Without the support of these organizations this research could not have been completed.
`
`A dissertation designed to demonstrate human body tracking using inertial/magnetic
`
`sensors would not be very creditable without any inertial/magnetic sensors. Though much
`
`of the time I had no idea what he was talking about, it was the skills and expertise of Doug
`
`McKinney that produced the first functioning MARG sensors.
`
`The period during which this research was conducted was marked by tremendous
`
`change in my life. Much of it was extremely difficult and I often felt that everything was
`
`falling apart. Ann Flood helped me keep all the pieces together. Geographically distant, but
`
`close nonetheless, the support of my sisters Gail and Lynn and my mother also helped me
`
`to keep those pieces together. To all my friends who patiently listened to my whining, I also
`
`give my thanks.
`
`Only my daughter, Carol, knows what it has been like to live me as I worked to finish
`
`this dissertation. She has put up with me and has been far more important to its completion
`
`than she will even know. I often think of her and wonder how one so young ever got to be
`
`so wise. Carol, I think the good times may finally be here.
`
`xvi
`
`
`
`DEDICATION
`
`For Carol and all for Carol
`
`xvii
`
`
`
`
`
`
`
`XViil
`xviii
`
`
`
`In memory of Robert G. Bachmann
`
`xix
`
`
`
`
`
`
`
`XX
`xx
`
`
`
`I. INTRODUCTION
`
`A.
`
`MOTIVATION
`
`Fully developed networked synthetic environments (SE) stand to revolutionize the
`
`fields of education, training, business, retailing and entertainment. They will fundamentally
`
`alter our societies and the way in which mankind views the world. In the educational field,
`
`synthetic environments will offer the ultimate in hands-on and visualization of difficult
`
`concepts. They will allow training to transpire in a place much like that in which the skills
`
`being practiced will be used without exposure to possible hazards and at less cost. In the
`
`workplace, employees will be able to work "side by side" even though they may be
`
`physically separated by hundreds or even thousands of miles. Using synthetic
`
`environments, corporations will obtain a safe, economical and efficient method of testing
`
`new concepts and systems. Retailers will create virtual department stores where consumers
`
`will be able to try out products to an unprecedented degree before actually buying them.
`
`Using synthetic environments, the entertainment industry will be able to create entire
`
`worlds in which customers will be able to experience thrills and live out entire fantasy lives.
`
`[Ref. 21.][Ref. 97.]
`
`The power of the synthetic environment lies in its ability to immerse users in a
`
`different world. The more complete the immersion, the more effective the synthetic
`
`environment. For complete immersion, the user should sense and interact with the synthetic
`
`environment in the same manner in which interaction with the natural world takes place.
`
`Interaction in the natural world results from body motion. Information regarding the
`
`surrounding environment is obtained through the five senses. Changes in body posture and
`
`position directly affect what is seen, heard, felt and smelled. The parameters sensed in the
`
`environment are altered and manipulated by the actions of the body. Thus, in order for a
`
`user to interact with a synthetic environment in a natural way and have the synthetic
`
`environment present appropriate information to the senses, it is imperative that data
`
`regarding body motion and posture be obtained. Body posture and location data are also
`
`1
`
`
`
`needed in multi-user environments to drive the animation of avatars which represent the
`
`actions of users of the environment to each other.
`
`At this time, there is no practical and intuitive interface that allows an individual
`
`human to be inserted into a SE in a fully immersive manner. Numerous motion tracking
`
`technologies are currently in use, but each suffers from its own set of limitations.
`
`Depending on the technology, these limitations may include marginal accuracy, user
`
`encumbrance, restricted range, susceptibility to interference and noise, poor registration,
`
`occlusion difficulties and high latency. Due to these problems, real-time animations of
`
`avatars must be largely script-based using motion libraries. For the most part, only a single
`
`u