Artificial neural networks applied to the classification of hand gestures using eletromyographic signals

Abstract

This paper aims at the classification of hand gestures using electromyographic signals (EMG) obtained through a Myo <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">TM</sup> armband, which has eight medical grade electrodes. Each electrode provides information regarding muscles contraction performed during the execution of the movement. From these electrodes signals are extracted seven features for each one of eight electrodes. After extraction of the characteristics a Feedforward Artificial Neural Network is trained to recognize the desired classes. The motivation of this research is the recognition of gestures for human-robot interaction. Experimental results are presented to demonstrate the performance of the proposed method.