A wavelet feature based mechanomyography classification system for a wearable rehabilitation system for the elderly

Abstract

This paper proposes a pattern recognition based system for identification of the forearm movements using Mechanomyography(MMG) for the rehabilitation of the elderly. The system is used to assist in the relearning and rehabilitation of the movements of the wrist and the hand. Surface MMG signals acquired from the flexor carpi ulnaris, brachioradialis supinator and abductor pollicis longus. The MMG is processed and wavelet based features are extracted which are classified into eight different forearm movements using a multilayer perceptron (MLP) classifier. A classification efficiency of 90.2 % is achieved using the MLP classifier. The MMG system is designed to measure data using accelerometers built into the assistive device and, hence, doesn't require any active involvement of the elderly. © 2013 Springer-Verlag Berlin Heidelberg.