A multi-robot path-planning algorithm for autonomous navigation using meta-reinforcement learning based on transfer learning

Wen, Shuhuan1,2; Wen, Zeteng1,2; Zhang, Di1,2; Zhang, Hong3; Wang, Tao1,2

2021-10-01

10.1016/j.asoc.2021.107605

APPLIED SOFT COMPUTING   影响因子和分区

1568-4946

1872-9681

The adaptability of multi-robot systems in complex environments is a hot topic. Aiming at static and dynamic obstacles in complex environments, this paper presents dynamic proximal meta policy optimization with covariance matrix adaptation evolutionary strategies (dynamic-PMPO-CMA) to avoid obstacles and realize autonomous navigation. Firstly, we propose dynamic proximal policy optimization with covariance matrix adaptation evolutionary strategies (dynamic-PPO-CMA) based on original proximal policy optimization (PPO) to obtain a valid policy of obstacles avoidance. The simulation results show that the proposed dynamic-PPO-CMA can avoid obstacles and reach the designated target position successfully. Secondly, in order to improve the adaptability of multi-robot systems in different environments, we integrate meta-learning with dynamic-PPO-CMA to form the dynamic-PMPO-CMA algorithm. In training process, we use the proposed dynamic-PMPO-CMA to train robots to learn multi-task policy. Finally, in testing process, transfer learning is introduced to the proposed dynamic-PMPO-CMA algorithm. The trained parameters of meta policy are transferred to new environments and regarded as the initial parameters. The simulation results show that the proposed algorithm can have faster convergence rate and arrive the destination more quickly than PPO, PMPO and dynamic-PPO-CMA. (C) 2021 Elsevier B.V. All rights reserved.

Multi-robot system Path planning Deep reinforcement learning Meta learning Transfer learning

SCI ; EI

英语

其他

National Natural Science Foun-dation of China[61773333] ; Royal Society of the United Kingdom Cooperation and Exchange Project[62111530148]

Computer Science

Computer Science, Artificial Intelligence ; Computer Science, Interdisciplinary Applications

WOS:000713098900001

ELSEVIER

COMPUTER SCIENCE

Web of Science

1.Yanshan Univ, Engn Res Ctr, Minist Educ Intelligent Control Syst & Intelligen, Qinhuangdao 066004, Hebei, Peoples R China
2.Yanshan Univ, Key Lab Ind Comp Control Engn Hebei Prov, Qinhuangdao 066004, Hebei, Peoples R China
3.Southern Univ Sci & Technol, Dept Elect & Elect Engn, Shenzhen 518000, Peoples R China

Wen, Shuhuan,Wen, Zeteng,Zhang, Di,et al. A multi-robot path-planning algorithm for autonomous navigation using meta-reinforcement learning based on transfer learning[J]. APPLIED SOFT COMPUTING,2021,110.

Wen, Shuhuan,Wen, Zeteng,Zhang, Di,Zhang, Hong,&Wang, Tao.(2021).A multi-robot path-planning algorithm for autonomous navigation using meta-reinforcement learning based on transfer learning.APPLIED SOFT COMPUTING,110.

Wen, Shuhuan,et al."A multi-robot path-planning algorithm for autonomous navigation using meta-reinforcement learning based on transfer learning".APPLIED SOFT COMPUTING 110(2021).