关键词: 聚合经验模态分解, 本征模态函数, 模糊隶属度, 噪声主导模态, 信号去噪

Abstract:

In order to improve the denoising effect of the noise dominant mode in EEMD decomposition, we propose a novel denoising method combining the EEMD and fuzzy threshold by using fuzzy membership degree. Firstly, the similarity between the intrinsic density function (IMF) and the probability density function (PDF) of the observed signals is calculated using the two norms, and the noise-dominated IMF is obtained. Then, the noise-dominated IMF is subjected to fuzzy threshold processing and hence the noise is removed from the IMF. Finally, all of the remained IMFs are reconstructed to get noise suppression signals. Simulation experiments are conducted by using both suppositional and ECG signals. The results show that the denoising effect of the proposed method is better than that of the wavelet half-soft threshold method and the EMD-based interval threshold (EMD-IT) method.

Key words: ensemble empirical mode decomposition, intrinsic mode function, fuzzy membership degree, noise dominant mode, signal denoising