Actuation of prosthetic drive using EMG signal

Myoelectric or electromyogram (EMG) signals can be useful in intelligently recognizing intended limb motion of a person. This paper presents an attempt to develop a four-channel EMG signal acquisition system as part of an ongoing research in the development of an active prosthetic device. The acquired signals are used for identification and classification of six unique movements of hand and wrist, viz. hand open, hand close, wrist flexion, wrist extension, ulnar deviation and radial deviation. This information is used for actuation of prosthetic drive. The time domain features are extracted, and their dimension is reduced using principal component analysis. The reduced features are classified using two different techniques: k nearest neighbor and artificial neural networks, and the results are compared.