1P2-K05 Training System for the Five-finger EMG-prosthetic Hand Using Virtual Reality

Abstract

This paper proposes a training system for controlling the five-finger EMG-prosthetic hand that enables not only fast finger motions but also a strong grasping force. The proposed system is developed based on the control method equivalent to the five-finger prosthetic hand, and conducts three types of training: muscle contraction, muscle cooperation, and grip force training. In the contraction training and the cooperation training, the control ability of muscle force and the reproducibility of EMG patterns can be trained, respectively. Appropriate gripping force according to the hardness of an object can also be trained via the grip force training using a virtual prosthetic hand. The validity of the proposed system is demonstrated in the five-day experiments conducted for three subjects.