The Mobile Autonomous Robot System (MARS) project is a vision based navigational system for a mobile platform. This thesis covers the design of image acquisition, image information extraction (vision) - using spectral pattern matching, obstacle map building, simple navigation, and integration into a working platform. The hardware design and implementation for both the platform and the vision system are also presented here. The thesis concludes with examples of test runs, and evaluation of overall system performance.

This paper is dedicated to the Mars Observer spacecraft that self destructed somewhere in the vicinity of Mars at 6pm Pacific Daylight Time, 21st Aug 1993. Rust In Peace.

I would like to thank my thesis supervisor Claude Sammut for his help and ideas during the development of the thesis. Much credit also goes to my parents who have been very supportive throughout the entire project and their ``Get back to work!'' words of encouragement. I wish to thank my sister Ivana and Eric McCreath for proof weeding the thesis. I would also like to thank my friends James Duff, Eric McCreath, Helmut Hirsh, Greg Konstantinidis, Wayne Earnshaw, Ross Mark, Amanda Chandler, Karen Luu, and the staff of the computer Science and Engineering Schools for their great support and for putting up with me when I started raving on and on.