Modified Q-learning with distance metric and virtual target on path planning of mobile robot

作者:

Highlights:

• We propose an improved Q-learning for path planning of mobile robot.

• The ideas of distance metric, moving target and modified Q function is integrated.

• The effectiveness of the improved Q-learning is tested in twenty navigation maps.

• The proposed improved Q-learning outperforms Q-learning.

• The performance of improved Q-learning is comparable with other path planners.

摘要

•We propose an improved Q-learning for path planning of mobile robot.•The ideas of distance metric, moving target and modified Q function is integrated.•The effectiveness of the improved Q-learning is tested in twenty navigation maps.•The proposed improved Q-learning outperforms Q-learning.•The performance of improved Q-learning is comparable with other path planners.

论文关键词:Moving target,Obstacle avoidance,Path planning,Q-learning, reinforcement learning,Mobile robot

论文评审过程:Received 19 June 2020, Revised 15 March 2022, Accepted 2 April 2022, Available online 7 April 2022, Version of Record 13 April 2022.

论文官网地址:https://doi.org/10.1016/j.eswa.2022.117191