[15p-M_123-1~8] FS.1 Focused Session "AI Electronics"

4:00 PM - 4:15 PM JST(7:00 AM - 7:15 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[15p-M_123-8] Improvement of Controllability of Myoelectric Prosthesis Hand Based on Reservoir Computing Framework (2)

〇Yusuke Hoshika¹, Zenji Yatabe¹, Seiya Kasai¹ (1.RCIQE, Hokkaido Univ.)

Keywords:

Myoelectric prosthesis hand,Surface myoelectric signal analysis,Reservoir computing

To achieve controllable myoelectric prostheses, we have developed a unique machine-learning-based myoelectric signal analysis technique considering the similarity between reservoir computing (RC) and the human motor control system. In this paper, we figure out and solve the problem of our technique, in which a part of hand movements such as supinator cannot be responded. From the careful observation and comparison between the activated supination and the myoelectric signals taken on a forearm, we find that the supination is mainly made by the muscle in the upper arm, not the forearm. Based on this understanding, we modify the motion of the supination to activate the muscle in the forearm and we succeed in the detection of the myoelectric signal and achieve correct inference of the supination, which demonstrate the feasibility of our technique.