Behavior Based Control and Fuzzy Q-learning for Autonomous Mobile Robot Navigation

This paper presents collaboration of behavior based control and fuzzy Q-learning for mobile robot navigation systems. There are many fuzzy Qlearning algorithms that have been proposed to yield individual behavior like obstacle avoidance, find target and so on. However, for complicated tasks, it is needed to combine all behaviors in one control schema using behavior based control. Based this fact, this paper proposes a control schema that incorporate fuzzy q-learning in behavior based schema to overcome complicated tasks in navigation systems of autonomous mobile robot. In the proposed schema, there are two behaviors which is learned by fuzzy q-learning. Other behaviors is constructed in design step. All behaviors are coordinated by hierarchical hybrid coordination node. Simulation results demonstrate that the robot with proposed schema is able to learn the right policy, to avoid obstacle and to find the target. However, Fuzzy q-learning failed to give right policy for the robot to avoid collision in the corner location.