Hybridization of fuzzy Q-learning and behavior-based control for autonomous mobile robot navigation in cluttered environment

This paper proposes hybridization of fuzzy Q-learning and behavior-based control for autonomous mobile robot navigation problem in cluttered environment with unknown target position. The fuzzy Q-learning is incorporated in behavior-based control structure and it is considered as generation of primitive behavior like obstacle avoidance and target searching. The simulation result demonstrates that the hybridization enables robot to be able to learn the right policy, to avoid obstacle and to find the target. Real implementation of this hybridization shows that the robot was able to learn the right policy i.e. to avoid obstacle.