Human support robot capable of flexible decision-making based on environment and self-functional assessment

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K14102
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 20010:Mechanics and mechatronics-related
• Research Institution: University of Tsukuba
• Principal Investigator: Ayanori Yorozu 筑波大学, システム情報系, 助教 (40781159)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: 行動・動作計画 / 移動マニピュレータ型ロボット / 未知環境探索 / 掃除計画 / 集合被覆問題

Outline of Final Research Achievements

In order for a robot to perform complex tasks in a human living environment, it must be able to select the necessary actions and behaviors as appropriate to achieve the target state and execute them based on its own functions and surrounding environments. For a 3D environmental mapping by a mobile manipulator-type robot equipped with an RGB-D camera at its tip, we proposed a method to sequentially determine the movement and camera posture, taking into account the observation range and the amount of unobserved area that can be observed. It was confirmed that the proposed method can realize efficient search behavior. For wiping and cleaning tables and shelves, we also showed that efficient work can be performed in terms of path length and work time by applying the set covering problem, considering the range reached by the tip of the manipulator and the area to be cleaned.

Free Research Field

知能ロボティクス

Academic Significance and Societal Importance of the Research Achievements

生活支援ロボットの研究が盛んに行われ、要求されるタスクも複雑になりつつある。本研究では、移動マニピュレータ型ロボットを対象に、移動・マニピュレーション・観測の機能の観点から行動・動作計画を行う方法を提案した。3次元環境地図生成において、移動とカメラ姿勢を逐次決定することで、効率的な3次元環境探索手法を提案した。また、テーブルや棚の拭き掃除に対して、マニピュレータの先端が届く範囲と掃除対象領域を考慮して、集合被覆問題を適用することで、経路長と作業時間の観点で効率的な掃除動作計画を提案した。提案のフレームワークは拭き作業だけでなく、探索や片付けなど生活環境で求められる作業への応用も可能である。