User-guided motion planning with reinforcement learning for human-robot collaboration in smart manufacturing

作者：

Highlights：

• Develop a method to abstract kinematic features of human demonstrations and tasks.

• Propose criteria to identify semantic similarity between tasks.

• Develop a mapping method to enable robots to learn from human demonstrations.

• Formulate the LfD problem as an MDP and solve the problem using Q-learning.

摘要

•Develop a method to abstract kinematic features of human demonstrations and tasks.•Propose criteria to identify semantic similarity between tasks.•Develop a mapping method to enable robots to learn from human demonstrations.•Formulate the LfD problem as an MDP and solve the problem using Q-learning.

论文关键词：Human-robot collaboration,Learning from demonstration,Motion planning,Reinforcement learning

论文评审过程：Received 21 February 2022, Revised 11 July 2022, Accepted 25 July 2022, Available online 6 August 2022, Version of Record 11 August 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118291