2022 Fiscal Year Final Research Report
Latent feature evaluation of muscle fatigue based on scale mixture stochastic model and its application to adaptive control of myoelectric prosthetic hand
Project/Area Number |
20K14698
|
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 | Hiroshima University |
Principal Investigator |
Furui Akira 広島大学, 先進理工系科学研究科(工), 助教 (30868237)
|
Project Period (FY) |
2020-04-01 – 2023-03-31
|
Keywords | 筋電位信号 / 生体信号 / 機械学習 / 確率モデル / 筋疲労 / 逐次学習 / 筋電義手 |
Outline of Final Research Achievements |
This study developed a methodology for inferring the latent state of electromyogram (EMG) signals during muscle fatigue using a stochastic model based on scale-mixture representation. Secondarily, we showed that this methodology can be applied to biological signals other than EMG signals, such as electroencephalography (EEG). We also extended the model and proposed an EMG pattern classifier that can account for uncertainty during muscle activity and an adaptive motion recognition method based on Bayesian sequential learning. Furthermore, we developed a myoelectric prosthetic hand that can realize biomimetic movements by introducing the proposed classifiers into a prosthetic hand control system.
|
Free Research Field |
生体信号解析,機械学習,確率モデル
|
Academic Significance and Societal Importance of the Research Achievements |
本研究成果の学術的意義は,筋電位信号の時系列データに含まれる潜在的特徴を推定可能な,新たな確率モデリングの枠組みを提案している点である.これに加え,モデルをパターン分類法へと展開し,筋疲労に対して頑健な動作認識に応用したことも,本研究の貢献として挙げられる. 筋電位信号などの生体信号から,ヒトの動作意図や異常の兆候といった内在的な情報を推定することができれば,より直感的かつ自然に操作可能なヒューマンマシンインタフェースに繋げることができる.本研究は,そのような技術を実現する上で考慮しなければならない「筋疲労」という現象の対処に焦点を当てたものであり,この点に本研究成果の社会的意義がある.
|