Maximising Coefficiency of Human-Robot Handovers through Reinforcement Learning

Journal: IEEE Robotics and Automation Letters (RA-L) | Paper link: doi.org/10.1109/LRA.2023.3280752 | arXiv preprint: arxiv.org/abs/2306.07205
This work proposes a novel human-robot interaction framework where the cognitive science principle that "humans act coefficiently as a group" (i.e. simultaneously maximising the benefits and minimising the effort of all agents involved) was transferred into human-robot cooperative settings. We modelled a human-robot coefficiency score by online capturing implicit human comfort and discomfort body signals as well as the robot expenses. This score is used as reward in a reinforcement learning (RL) problem to adapt and learn online the combination of robot interaction parameters that maximises such coefficiency.
This work was supported by the ERC-StG Ergo-Lean (Grant Agreement No.850932) and The Royal Society (Grant Agreement No. IES/R3/203086).

[@fawadkhan8905]

Can we have the code please??