Double Q-PID algorithm for mobile robot control

作者：

Highlights：

• A model-free reinforcement learning algorithm for adaptive low-level PID control.

• Double Q-learning provides excelent performance with low computational cost.

• Incremental state and action spaces discretization for fast reinforcement learning.

• The Double Q-PID algorithm is effectively applied to multiple robotic platforms.

• The proposed algorithm effectively improves the adaptability of the PID controllers.

摘要

•A model-free reinforcement learning algorithm for adaptive low-level PID control.•Double Q-learning provides excelent performance with low computational cost.•Incremental state and action spaces discretization for fast reinforcement learning.•The Double Q-PID algorithm is effectively applied to multiple robotic platforms.•The proposed algorithm effectively improves the adaptability of the PID controllers.

论文关键词：Reinforcement learning,Double Q-learning,Incremental learning,Double Q-PID,Mobile robots,Multi-platforms

论文评审过程：Received 1 November 2018, Revised 26 June 2019, Accepted 26 June 2019, Available online 27 June 2019, Version of Record 9 July 2019.

论文官网地址：https://doi.org/10.1016/j.eswa.2019.06.066