Multichannel surface electromyography classification based on muscular synergy

Abstract

With the aim to control a multiple degrees of freedom electromechanical devices, e.g., assistive robots, powered wheelchair, etc., this paper proposes a real-time multichannel surface electromyography classification scheme based on the coordination or synergies between a functional group of muscles: biceps brachii, triceps brachii, pronator teres, and brachioradialis. The muscular synergy is evaluated by the analysis of a multivariate function, composed by the four corresponding neuromuscular activation functions, and the cross-correlation matrix of muscular force estimated through the root mean square (RMS) value of sEMG amplitude. The resulting features from the training set were used to train an artificial neural network with classification accuracy up of 90%.