Path planning method for obstacle avoidance mobile robot

The invention relates to a robot path planning method, in particular to a path planning method for an obstacle avoidance mobile robot, which comprises a calculation process and a training process. Firstly, a mobile robot path planning algorithm based on deep reinforcement learning in a sparse environment is proposed, and an exploration strategy adopts a confidence level to realize balance between exploration and utilization; through the use of a post-view playback experience mechanism, the problem that convergence is difficult in a sparse reward environment is solved, the sample utilization rate is improved, the convergence speed is increased, and the situation that reinforcement learning needs to design a complex reward function in path planning is avoided to a certain extent. According to the method, the executed action is selected through the confidence interval upper bound exploration strategy, a post experience playback mechanism is adopted, the arrived state serves as a target, the mobile robot can obta