Towards autonomous mobile robot navigation

Abstract

THE UMASS MOBILE ROBOT PROJECT IS INVESTIGATING THE PROBLEM OF INTELLI- GENT NAVIGATION OF AN AUTONOMOUS ROBOT VEHICLE. MODEL-BASED PROCESSING OF THE VISUAL SENSORY DATA IS THE PRIMARY MECHANISM USED FOR OBSTACLE AVOID- ANCE, MOVEMENT THROUGH THE ENVIRONMENT, AND MEASURING PROGRESS TOWARD A GIVEN GOAL. THIS PAPER DESCRIBES OUR CURRENT APPROACH TO GOAL-ORIENTED NAVIGATION THROUGH A PARTIALLY MODELED, UNCHANGING ENVIRONMENT WHICH CONTAINS NO UNMODELLED OBSTACLES. THE NAVIGATION SYSTEM INTEGRATES PERCEPTION, PLANNING, AND EXECUTION OF ACTIONS. OF PARTICULAR IMPORTANCE IS THAT THE PLANNING PROCESSES ARE ABLE TO REASON ABOUT LANDMARKS THAT SHOULD BE PERCEIVED AT VARIOUS STAGES OF PLAN EXECUTION. EXPERIMENTS IN THIS AND THREE COMPANION PAPERS DEMONSTRATE THE PERFORMANCE OF THE VARIOUS COMPONENTS WITHIN THE NAVIGATION SYSTEM.