AWGN Suppression Algorithm in EMG Signals Using Ensemble Empirical Mode Decomposition

Surface Electromyogram (EMG) signals are often contaminated by background interferences or noises, imposing difficulties for myoelectric control. Among these, a major concern is the effective suppression of Additive White Gaussian Noise (AWGN), whose spectral components coincide with the spectrum of EMG signals; making its analysis problematic. This paper presents an algorithm for the minimization of AWGN from the EMG signal using Ensemble Empirical Mode Decomposition (EEMD). In this methodology, EEMD is first applied on the corrupted EMG signals to decompose them into various Intrinsic Mode Functions (IMFs) followed by Morphological Filtering. Herein, a square-shaped structuring element is employed for requisite filtering of each of the IMFs. The outcomes of the proposed methodology are found improved when compared with those of conventional EMD-and EEMD-based approaches.