Signal Processing of Surface Detected Electromyographic Activity

Abstract

<p>A technique is presented for assessing the relative performance of various signal processors of the surface detected electromyographic signal (EMG) in the gross skeletal muscles of man. A minicomputer is used to sample, store, and to later process the EMGs for agonists of the upper arm for various net forces, as measured at the wrist, under a condition of isometric tension. A two dimensional analysis of the flexor and extensor EMGs is performed for each force level. The number of force levels from which distinct, high confidence, control signals may be derived is used as a figure of merit to determine the superior of the signal processors studied and the superior of various electrode sites considered. For prosthetic use it is often desirable to maximize the number of control signals per muscle site.</p>