A neural network-based surface electromyography motion pattern classifier for the control of prostheses

This paper presents a surface electromyography (EMG) motion pattern classifier which combines an artificial neural network (ANN) with a parametric model such as an autoregressive (AR) model. This motion pattern classifier can successfully identify four types of movement of human hand, wrist flexion, wrist extension, forearm pronation and forearm supination, by using the surface EMG detected from the flexor carpi radialis and the extensor carpi ulnaris. This desirable result shows that it have a great potential application to our Tsinghua multi-degree artificial hand.