Variational mode decomposition based on refined composite multiscale dispersion entropy and its application in sEMG...

Area of Science:

• Biomedical Engineering
• Signal Processing
• Machine Learning

Background:

• Surface electromyography (sEMG) signals are inherently noisy and non-stationary, complicating analysis.
• Existing denoising methods often struggle with low signal-to-noise ratio (SNR) conditions due to suboptimal mode selection.
• These limitations hinder the widespread application of sEMG in various fields.

Purpose of the Study:

• To develop an advanced hybrid denoising framework for sEMG signals.
• To improve the identification and suppression of noise-dominant components in sEMG.
• To enhance the performance of sEMG-based applications, particularly under low SNR.

Main Methods:

• Proposed a hybrid framework integrating Variational Mode Decomposition (VMD) and Refined Composite Multiscale Dispersion Entropy (RCMDE).
• Developed an RCMDE-weighted VMF screening mechanism for effective noise component identification and spectral leakage suppression.
• Combined Local Mean Decomposition (LMD) and Wavelet Thresholding (WT) for superior denoising.

Main Results:

• The proposed algorithm demonstrated superior performance in terms of SNR and RMSE compared to conventional VMD and other methods.
• Significant improvements were observed particularly in low-SNR scenarios.
• Gesture classification accuracy was enhanced using the denoised sEMG signals.

Conclusions:

• The hybrid VMD-RCMDE framework offers an efficient and effective preprocessing solution for sEMG signals.
• This method significantly improves denoising performance, especially under challenging low-SNR conditions.
• The findings support broader applications of sEMG in medical diagnosis and human-computer interaction.