Low DOF representation of hand movements and precise control of a multi-fingered robot hand based on muscle synergies

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K14125
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 20020:Robotics and intelligent system-related
• Research Institution: Tokyo University of Science (2023); The University of Electro-Communications (2021-2022)
• Principal Investigator: Yamanoi Yusuke 東京理科大学, 工学部電気工学科, 助教 (40870184)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: 筋電義手 / 運動機能解析

Outline of Final Research Achievements

This study proposes a novel control method for myoelectric prosthetic hands. Electromyogram (EMG) signals are biosignals that occur by muscular contractions and it is possible to extract the intention to move the hand from amputees. Therefore, the prosthetic hand can be controlled by the wearer's intention.
However, current control methods achieve only simple movements of the prosthetic hand compared with a healthy hand because EMG signals are very weak and fragile and it is difficult to extract the intention in detail.
For this issue, this study analysed the movement of hands by using a cutting-edge machine learning method and an image-recognition-type motion capture and we developed the control method for myoelectric prosthetic hands to drive each finger independently.

Free Research Field

生体信号認識

Academic Significance and Societal Importance of the Research Achievements

本研究では筋電センサと画像認識型モーションキャプチャを組み合わせて，複数の被験者からデータ収集を行いデータセットを構築した．手は身体の中で最も関節が密集している部位であり，全部の関節にセンサを取り付けて関節角度を測るということは不可能であったが，近年の画像認識技術の向上によって非接触に計測することができるようになった．今回構築したデータセットはヒトの筋活動と手指運動の関係性の解明，義手の制御手法にとって非常に有用である．現在，研究レベルでは手指の動作パターンを認識し制御する義手が主流であるが，本研究によって各指の詳細な関節角度を推測し制御する手法を構築することができた．