ABSTRACT

The Center for Self-Organizing and Intelligent Systems (CSOIS) at Utah State University has been developing laser-based obstacle detection and avoidance during the past four years. The Jet Propulsion Laboratory (JPL) initially encouraged us down this path to provide an optional software upgrade for its Rocky Rover. The USU system was based on a unique combination of fuzzy logic and a decision tree behavioral technique, and this system provided the Rocky Rover the option of venturing independently beyond the observance of the base station. Shortcomings in the data collection prompted a redesign that included more sophisticated line tracing and filtering methods to locate and track the actual laser lines. The signal to noise ratio problems combined with platform changes lead to the choice of a commercial range finder. The design challenge then became one of system integration including the meeting of real time constraints on the current processor platform, and obstacle characterization and avoidance.

Read Entire Publication PDF