HIERARCHICAL LEARNING OF REACTIVE BEHAVIORS IN AN AUTONOMOUS MOBILE ROBOT

Abstract. We describe an autonomous mobile robot that employs a simple sensorimotor learning algoríthm at three different behavioral levels to achieve coherent goal-directed behavior. Therobotautonomously navigates to a goal destination within an obstacle-ridden environment by using the learned behaviors of obstacle detection, obstacle avoidance, and beacon following. These reactive behaviors are learned in a hierarchical manner by using a simple hillclimbing routine tbat attempts t~ find tbe optimal transfer function from perceptions to actions for each behavior. We present experimental results which show that each behavior was successfully learned by the robot witbin a reasonably short period oftime. We conclude by discussing salient features of our approach and possible directions for future research