APPLICATION OF NEURAL NETWORKS TO MOTION PLANNING AND CONTROL FOR AN ARTICULATED ROBOT ARM

by Hugh Jack

Department of Mechanical Engineering Faculty of Engineering Science

Submitted in partial fulfillment of the requirements for the degree of Master of Engineering Science

Faculty of Graduate Studies, The University of Western Ontario, London, Ontario, January, 1991

© Hugh Jack 1991

THE UNIVERSITY OF WESTERN ONTARIO FACULTY OF GRADUATE STUDIES

CERTIFICATE OF EXAMINATION

1.1 ABSTRACT

1.2 AKNOWLEDGEMENTS

1.3 TABLE OF CONTENTS

1.4 LIST OF APPENDICES

1.5 LIST OF FIGURES

1.6 GLOSSARY OF TERMS

1. INTRODUCTION

1. A REVIEW OF ROBOT MOTION PLANNING

CHAPTER 3 - ARTIFICIAL NEURAL NETWORKS

CHAPTER 4 - A SURVEY OF NEURAL NETWORK APPLICATIONS IN ROBOTICS

CHAPTER 5 - A SIMPLE ROBOT MODEL

CHAPTER 6 - FEED FORWARD NEURAL NETWORKS FOR MOTION PLANNING

CHAPTER 7 - MAXIMUM VELOCITY MOTION PLANNER

CHAPTER 8 - MAXIMUM ACCELERATION MOTION PLANNER

CHAPTER 9 - SPLINES FOR MOTION MODELLING

CHAPTER 10 - FINDING NEAR OPTIMAL ROBOT MOTIONS

CHAPTER 11 - NEAR-OPTIMAL MOTION WITH BEZIER SPLINES

CHAPTER 12 - MAXIMUM TORQUE PATH PLANNING

CHAPTER 13 - MAXIMUM TORQUE WITH SCALED NEURAL NETWORK

CHAPTER 14 - DISCUSION AND CONCLUSION

APPENDIX A - Kinematics of a Two Link Revolute Manipulator

APPENDIX B - DYNAMICS OF TWO LINK MANIPULATORS

APPENDIX C - NEURAL NETWORK SUBROUTINES

APPENDIX D - ADAPTIVE TIME STEP RUNGE-KUTTA INTEGRATION

APPENDIX E - GLOBAL DEFINITIONS FOR SOFTWARE

APPENDIX F - DYNAMICS MODULE

APPENDIX G - HOOKES AND JEEVES OPTIMIZATION MODULE

APPENDIX H - PATH OPTIMIZATION SUBROUTINES

APPENDIX I - NEURAL NETWORK MOTION PLANNING MODULE

APPENDIX J - PATH PLANNING FUNCTION MODULE

APPENDIX K - NEURAL NETWORK BATCH TRAINING PROGRAM

APPENDIX L - USER INTEFACE PROGRAM

REFERENCES