基于CM-Q学习的自主移动机器人局部路径规划

Local path planning of autonomous mobile robot based on CM-Q learning

针对多障碍物未知环境下,自主移动机器人局部路径规划过程中出现的路径冗余和避障问题,提出了基于坐标匹配的Q学习算法(Coordinate Matching-Q learning算法,CM-Q算法)。首先建立自主移动机器人栅格地图运行环境;其次以Q学习算法探索和学习最佳状态-动作对,并利用坐标匹配的CM算法进行避障;最后在未知障碍物环境中进行路径规划,对所提出的算法进行验证。实验结果表明,运用该方法,自主移动机器人能在未知的简单和复杂障碍物环境下规划出一条最优或次优路径,完成避障和路径规划任务。

A coordinate matching Q-learning(CM-Q learning) algorithm is proposed to solve the problem of path redundancy and obstacle avoidance in the local path planning of the autonomous mobile robot under multiple obstacle unknown environment. Firstly, the grid map is established for the workplace of the autonomous mobile robot. Secondly, the Q-learning algorithm is used to explore and learn the best state-action pairs, and the CM algorithm is taken for obstacle avoidance. Finally, the proposed CM-Q algorithm is tested and verified by the mobile robot under unknown environment. The experimental results show that the autonomous mobile robot can plan an optimal or sub-optimal path under unknown simple or complex environments to achieve the tasks of obstacles′ avoidance and the local path planning.