Abstract:
We discuss six labs that present two contrasting approaches to AI robotics. The first four labs present the traditional, representation-based approach to designing behavior control algorithms for autonomous robots. Robots use an internal representation of the world plus a planning algorithm to guide behavior. The second two labs present a behavior-based or reactive approach to managing robot behavior. Behavior is tied directly to sensor readings, eliminating the need for an internal representation of the world or a centralized planner.