Emergence of idle robots in a multi-robot system and its verification by long-term experiments

2023 Fiscal Year Final Research Report

• Project/Area Number: 20K19902
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 61050:Intelligent robotics-related
• Research Institution: Wakayama National College of Technology
• Principal Investigator: Murayama Toru 和歌山工業高等専門学校, 知能機械工学科, 准教授 (30709176)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Keywords: 群ロボット / 耐故障性 / ネットワーク構造

Outline of Final Research Achievements

In this study, a control method to improve the fault-tolerance of network structures is developed for multi-robot systems in which robots are networked by wireless communication. In tasks where the coverage of sensors mounted by each robot is maximized, the network connectivity is disrupted by the failure of a small number of robots due to the sparse configuration of robots. In response to this problem, this study shows that the fault-tolerance of the entire system can be significantly improved if a small number of robots play a role that contributes to fault-tolerance instead of contributing to the task. An autonomous decentralized control law was devised to achieve such a behavior that contributes to fault-tolerance. The effectiveness of the proposed method was confirmed in long-duration tasks that require fault-tolerance and charging.

Free Research Field

ロボティクス

Academic Significance and Societal Importance of the Research Achievements

近年、ロボットは労働力の代替として期待されており、長い時間の運用が求められてきている。自律移動ロボットを長時間運用する際、故障のような突発的で予測できない停止や、バッテリーの充電のような一定期間ごとの停止は避けられない。本研究では群ロボットを対象に、耐故障性とタスク達成率のトレードオフを陽に考慮し、少数のロボットが耐故障性の役割を担うことで集団全体の耐故障性能とタスク達成を両立できることを明らかにした。さらに、耐故障性のあるネットワーク構造を自律的に構成するための自律分散制御手法を開発することで、群ロボットの実応用・社会実装へ近づけた。