|Robot Search Path Planning Method Based on Prioritized Deep Reinforcement Learning
Yanglong Liu, Zuguo Chen*, Yonggang Li, Ming Lu, Chaoyang Chen, and Xuzhuo Zhang
International Journal of Control, Automation, and Systems, vol. 20, no. 8, pp.2669-2680, 2022
Abstract : "The path planning process of the robot relies too much on environmental information, which makes it difficult to obtain the optimal search path when the search and rescue tasks are carried out in a complex postdisaster environment. Thus, a path planning method based on prioritized deep reinforcement learning is proposed in the paper. The core idea of the method is that the robot first builds an environment mathematical model based on the obtained information through the sensors. Then, to make the robot can obtain the optimal search policy in an extremely complex environment, the prioritized replay mechanism is used to improve deep reinforcement learning. Finally, the simulation results show that the search path planning method based on prioritized deep reinforcement learning proposed can not only improve the convergence speed of the model but also is endowed good robustness in this paper.
"Deep reinforcement learning, path planning, prioritized replay mechanism, robot search. "