Experimental Investigation of Electromyographic Activities of Upper Limb Muscles without and with a Passive Exoskeleton with Four Degrees of Freedoms

Abstract

The choice of suitable human surface muscles and of their electromyographic (EMG) signal processing have always been a challenge, especially when technical devices are to be controlled by these signals. Experiments with six healthy volunteers were performed with a NORAXON measuring and processing system. EMG signals from 8 muscles of the upper right arm performing the main motions in the shoulder and elbow joints were registered and processed. Four angles in the joints were also registered. Six motor tasks were performed, without and with an exoskeleton with four degrees of freedom. Since the EMG signal of the muscle pectoralis major was contaminated by noise coming from the electrical activity of the heart, first this signal was filtered using recently developed dynamic filter. Normalization according to maximal isometric tasks was performed after filtration. Rectification and smoothing ensured a suitable signal for proportional control of the exoskeleton. On/off control was simulated by setting different constants for maximal EMG amplitude levels.