Development of a Wearable and Dry sEMG Electrode System for Decoding of Human Hand Configurations

Abstract

Since bio-electric signals such as surface EMG are easily influenced by undesired artifacts and experimental environments including various electrical noises by peripheral devices, the amplifier is an issue of great importance. Although most commercial surface EMG amplifier systems provide high performance in acquiring electric bio-signals, they are not convenient for myoelectric control applications because they usually use wet-type electrodes that should be attached to the skin individually and there are also some limitations to possible modifications. In this study, we propose and develop a surface EMG interface that employs dry-type electrodes, a single supplied circuit for reduced weight, two voltage followers to improve input impedance, and a modified driven-right-leg circuit using a virtual ground circuit. By adapting a wearable band-type interface. The EMG electrodes can be reused while offering high performance corresponding to that of commercial products. The developed surface EMG system was successfully applied to decode human motion intentions of eight different configurations and a rest condition by using a fast training algorithms in a non-targeted manner.