EMG Denoising Based on CEEMDAN-PE-WT Algorithm.

Raw Surface Electromyography (sEMG) generally contains baseline noise, random noise, power line interference and other noises. The performance of signal denoising is a prerequisite for sEMG feature extraction and recognition. However, traditional filtering methods may sacrifice some effective sEMG signals during denoising process. For improving the denoising effect and eliminate the modal aliasing problem in the Empirical Mode Decomposition (EMD) decomposition process, this paper proposes a hybrid denoising algorithm based on complete ensemble empirical mode decomposition with Adaptive Noise (CEEMDAN) and permutation entropy (PE) combined with wavelet threshold (WT). Firstly, CEEMDAN decomposition is performed on raw sEMG signals to calculate the PE value of each Intrinsic Mode Functions (IMF). Then, high-frequency (H-F) IMF components dominated by random noise is identified and wavelet threshold denoising is applied. Results show that EMG signals denoised by the proposed CEEMDAN-PE-WT algorithm perform a higher signal-to-noise ratio (SNR) and lower root mean square error (RMSE) compared with WT, EMD and CEEMDAN denoising methods.