A motion generating system for multi-fingered myoelectric hand

The purpose of this study is to develop a control system that can generate various motions for multi-fingered myoelectric hand. Surface electromyogram signal that includes the information on the motor nerve is suitable for the control of artificial hand. However, we cannot presume detailed finger motions only by analyzing the surface electromyogram signal. Therefore, we propose a motion generating system. This system has the capability that generates a target motion automatically by combining the characteristic of some motions. In this paper, we develop a motion generating system using the neural networks and investigate whether this system can generate target motions in simulations.