Adaptability-Enhancement Scheme Based on Active Information-Exploration in Real Space and Efficient Motion-Exploration in Virtual Space

2020 Fiscal Year Final Research Report

• Project/Area Number: 18KT0063
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 特設分野
• Research Field: Intensification of Artifact Systems
• Research Institution: Waseda University
• Principal Investigator: KAMEZAKI MITSUHIRO 早稲田大学, 理工学術院総合研究所(理工学研究所), 主任研究員(研究院准教授) (30468863)
• Project Period (FY): 2018-07-18 – 2021-03-31
• Keywords: 能動的情報探索 / 効率的動作探索 / 災害対応ロボット / 実空間 / 仮想空間

Outline of Final Research Achievements

Autonomous robots that can work flexibly and robustly even in unknown environments needs to be able to judge the success or failure of a given task and to seek alternative solutions when the task is judged to be unsuccessful. In addition, failure is not allowed in disaster response tasks, so it would be effective to acquire behaviors based on trial and error in a virtual space, rather than in a real space, where risks exist and time is required. In this study, we developed a system to enhance the adaptability of autonomous robots in environments where trial and error in a real environment is difficult and the physical properties of the target are unknown, on the basis of active information exploration in real space and motion exploration in virtual space.

Free Research Field

知能機械・システム

Academic Significance and Societal Importance of the Research Achievements

安全性と迅速性が求められる災害対応などに従事する自律型作業ロボットには，目前のタスクと自身の現能力を照らし合わせた「実行可能性の判断」だけでなく，タスクが遂行できるように自身の能力を向上させていく「実行可能動作の学習」を行える機能が不可欠といえる．本研究では，実空間での能動的情報取得と仮想空間での効率的動作探索を核とした可否判断の慎重性と動作調整の迅速性を両立する「適応性強化システム」のフレームワークを提案する点で学術的意義および社会的意義が大きいといえる．