A wavelet transfom-based analysis of surface EMG signals

Abstract

Surface electromyogram (EMG) signals are biopotential signals, which give information about the behavior of muscles. Surface EMG signals are formed as the summation of motor unit action potentials (MUAP) obtained from active motor units. In this paper a method, which facilitates the analysis of real EMG data recorded at several contraction levels with surface electrodes, is introduced. In this method, in order to decompose surface EMG signals into their MUAP's, three different algorithms are employed and their performances are compared Wavelet coefficients of the EMG signal whose noise components are filtered by wavelet transform are decomposed into its independent components by a statistical method, called Independent Component Analysis (ICA). Determination of MUAP's has been achieved using the extracted independent components.