Research on Multi-point Continuous Electromyography Control Algorithm and Its System for Multi-Finger Dexterous Hand

Abstract

As the application field of robots expands from industry to agriculture, medical rehabilitation and service industries, its working environment becomes dynamic and not good. The end effectors of robots also need to be more flexible to adapt to changing operating objects. Based on the existing research of our group, a multi-finger dexterous hand included two kinds of bending joints is designed which used human hand as the prototype and combined with the characteristics of human hand joint movement. The EMG signal is proposed as the control source to improve the diversity and flexibility of dexterous hand movement. In the link of EMG signal processing algorithm, EMG signal feature value of the independent and complete action is discretized to form a continuous characteristic parameter curve. And multi-point continuous myoelectric control which is used to characterize the action process can be realized. The real-time adjustment of the driving air pressure is realized by the control system. And the system can control the different action forms of multi-finger dexterous hand. Finally, the relative error of proposed control algorithm and its system was verified to be about 3.23% by the joint position control test.