Research on Dynamic Path Planning of Mobile Robot Based on Improved DDPG Algorithm

Aiming at the problems of low success rate and slow learning speed DDPG algorithm in path planning a mobile robot dynamic environment, an improved is designed. In this article, RAdam used to replace neural network optimizer DDPG, combined with curiosity improve convergence speed. Based on algorithm, priority experience replay added, transfer introduced training effect. Through ROS operating system Gazebo simulation software, environment established, are compared. For task robot, results show that increased by 21%, 90% compared original algorithm. It has good effect for robots continuous action space.