Decomposition and evaluation of SEMG for hand prostheses control

Abstract

The evaluation of electromyogram signals (EMG) has been at the core of prosthetic control applications in the last decade, as prosthetic hands have become a valuable replacement for amputees for regaining some of the capabilities of their missing hand. The use of EMG signals as an input in decision-making systems offers a potential possibility for hand movement classification in the field of health and technology. The EMG signal can be recorded either using noninvasive or invasive measurement methods, however for many applications of human-computer interaction, non-invasiveness is a must. So, in this paper, seven different daily hand activities were recorded using surface electromyogram (SEMG) signals and analyzed using a statistical analysis approach to examine muscle force relationship based on analysis of variance and principal component analysis for hand movement interpretation. Muscle classification for control applications is essential to carrying out these activities, and once an appropriate algorithm for signal analysis is ready, the features of the signal can be easily investigated prior designing a prosthetic hand.