Competitive physical interaction by reinforcement learning agents using intention estimation

Hiroki Noda, Satoshi Nishikawa, Ryuma Niiyama, Yasuo Kuniyoshi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The physical human-robot interaction (pHRI) research field is expected to contribute to competitive and cooperative human-robot tasks that involve force interactions. However, compared with human-human interactions, current pHRI approaches lack tactical considerations. Current approaches do not estimate intentions from human behavior and do not select policies that are appropriate for the opponent's changing policy. For this reason, we propose a reinforcement learning model that estimates the opponent's changing policy using time-series observations and expresses the agent's policy in a common latent space, referring to descriptions of tactics in open-skill sports. We verify the performance of the reinforcement learning agent using two novel physical and competitive environments, push-hand game and air-hockey. From this, we confirm that the latent space works properly for policy information because each latent variable that represents the machine agent's own policy and that of the opponent affects the behavior of the agent. Two latent variables can clearly express how the agent estimates the opponent's policy and decides its own policy.

Original languageEnglish
Title of host publication2021 30th IEEE International Conference on Robot and Human Interactive Communication, RO-MAN 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages649-656
Number of pages8
ISBN (Electronic)9781665404921
DOIs
Publication statusPublished - Aug 8 2021
Event30th IEEE International Conference on Robot and Human Interactive Communication, RO-MAN 2021 - Virtual, Vancouver, Canada
Duration: Aug 8 2021Aug 12 2021

Publication series

Name2021 30th IEEE International Conference on Robot and Human Interactive Communication, RO-MAN 2021

Conference

Conference30th IEEE International Conference on Robot and Human Interactive Communication, RO-MAN 2021
Country/TerritoryCanada
CityVirtual, Vancouver
Period8/8/218/12/21

All Science Journal Classification (ASJC) codes

  • Human-Computer Interaction
  • Communication
  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Competitive physical interaction by reinforcement learning agents using intention estimation'. Together they form a unique fingerprint.

Cite this