Obstacle avoidance of hexapod robots using fuzzy Q-learning

摘要

Safe and autonomous obstacle avoidance plays an important role in the navigation control of hexapod robots. In this paper, we combine the method of reinforcement learning with fuzzy control to achieve the autonomous obstacle avoidance for a hexapod robot in complex environments. A fuzzy Q-learning algorithm is first presented and an obstacle avoidance approach is proposed using the Fuzzy Q-learning algorithm regarding the specific requirements of the hexapod robot. Then, the proposed approach is implemented for a real hexapod robot system that uses ultrasonic sensors to detect the obstacles in an unknown environment and learns an optimal policy to avoid the obstacles. Several groups of experiments are carried out to verify the performance of the proposed approach.