Sharing Experience for Behavior Generation of Real Swarm Robot Systems Using Deep Reinforcement Learning

• Yasuda Toshiyuki

  University of Toyama

• Ohkura Kazuhiro

  Hiroshima University

Abstract

Swarm robotic systems (SRSs) are a type of multi-robot system in which robots operate without any form of centralized control. The typical design methodology for SRSs comprises a behavior-based approach, where the desired collective behavior is obtained manually by designing the behavior of individual robots in advance. In contrast, in an automatic design approach, a certain general methodology is adopted. This paper presents a deep reinforcement learning approach for collective behavior acquisition of SRSs. The swarm robots are expected to collect information in parallel and share their experience for accelerating their learning. We conducted real swarm robot experiments and evaluated the learning performance of the swarm in a scenario where the robots consecutively traveled between two landmarks.

Journal

• Journal of Robotics and Mechatronics

  Journal of Robotics and Mechatronics 31 (4), 520-525, 2019-08-20

  Fuji Technology Press Ltd.

Keywords

• swarm robotics
• reinforcement learning
• deep Q network
• experience sharing
• real robot

Details

• CRID: 1390282763133310592
• NII Article ID: 130007691495
• NII Book ID: AA10809998
• DOI: 10.20965/jrm.2019.p0520
• ISSN: 18838049
• ISSN: 09153942
• NDL BIB ID: 029873785
• Web Site: http://id.ndl.go.jp/bib/029873785
• Web Site: https://ndlsearch.ndl.go.jp/books/R000000004-I029873785
• Web Site: https://www.fujipress.jp/main/wp-content/themes/Fujipress/phyosetsu.php?ppno=ROBOT003100040001
• Text Lang: en
• Data Source:
  • JaLC
  • NDL
  • Crossref
  • CiNii Articles
  • KAKEN
• Abstract License Flag: Disallowed