FastICA-based noise reduction algorithm for surface electromyographic signals

Surface Electromyography (sEMG) signals is an important element in human-computer interaction, and significantly affect the quality of subsequent research and applications. This study introduces a noise reduction approach based on Fast Independent Component Analysis (FastICA) algorithm. Firstly, process of FastICA is outlined briefly, and its theoretical feasibility is discussed. Then, in order to estimate the noise reduction effectiveness, sEMG sinal of grasping action which is served as the signal source has been acquired by a physiological signal amplifier. Finally, FastICA algorithm is employed to obtain independent components. After analyzing Kurtosis value of each component, identified gauissian noise will be set to zero to complete the noise reduction process. The result demonstrates that the FastICA-based noise reduction algorithm can eliminate Gaussian noise components in sEMG signals.