An Evaluation on EMG-based Machine Learning Classification of Hand Movements Using Three Electrodes Arrangement on Forearm

Abstract

A lot of types of myoelectric prosthetic hands using surface electromyogram have been investigated and developed in recent years. To control the myoelectric prosthetic hands, it is required to develop a high classification rate system. We propose a method of electrode placement and pretreatment by placing the three electrodes at the mid-forearm in the form of an armband. Even though similar studies have been developed in the past, we investigate the arrangement of electrodes among a lot of measurement points. To reduce the number of measuring electrodes, we evaluate the effects of muscle potential measurements on the pattern recognition and the classification of the useful measurement points by fixing three electrodes arranged in the form of armbands using the proposed method.