Online remote control of a robotic hand configurations using sEMG signals on a forearm

Abstract

This study presents an online remote control of a robotic hand using sEMG signals on a forearm. Eight skin surface electrodes were mounted on a forearm to detect the sEMG signals that correspond to four hand configurations and rest condition. In order to enhance learning speed and performance of the classifier, a supervised feature extraction method and a fast learning classifier were proposed. The hand configurations classified from the sEMG signals were delivered to a remote side robot hand via TCP-IP protocols. The demonstration verified that human could control the remote side robot hand in real-time using his or her sEMG signals with less than 50 seconds recorded training data.