摘要
结合互补集合经验模态分解(CEEMD)和基于排列熵的信号随机性检测,提出了MEEMD方法。通过采用MEEMD方法将一个含躁信号分解为几个固有模态(IMFS),用软阈值函数来抑制高频固有模态的噪声,提高信号的信噪比(SNR)。对比该方法与基于EEMD和小波软阈值的联合去噪、基于CEEMD和小波软阈值联合去噪等方法得到的信噪比(SNR)和平均平方误差(MSE),发现基于MEEMD小波软阈值去噪方法的去噪效果较好。
This paper proposes a denoising method based on MEEMD and wavelet soft threshold function. Because the white noise added by the EEMD decomposition can not be completely neutralized,the white noise of adding positive and negative pairs is proposed and the CEEMD decomposition is obtained. MEEMD was proposed in combination with CEEMD and signal randomness detection based on permutation entropy. A manic signal is decomposed by MEEMD decomposition approach into several intrinsic mode( IMFS). Because of the high frequency noise and low frequency drift interference,we need to use the soft threshold function to suppress the high frequency intrinsic mode noise and improve signal-to-noise ratio( SNR) of signals. This method is used to test the simulation and real data. By comparing with the SNR value and mean square error( MSE) based on EEMD and the combination of wavelet soft threshold denoising,we find that the de-noising effect of wavelet soft threshold denoising method based on the joint CEEMD and wavelet soft threshold denoising is better.
作者
李薇
白艳萍
LI Wei , BAI Yanping(School of Science ,North University of China ,Taiyuan 030051 ,China)
出处
《重庆理工大学学报(自然科学)》
CAS
北大核心
2018年第5期189-198,共10页
Journal of Chongqing University of Technology:Natural Science
基金
国家自然科学基金资助项目(61774137)
山西省自然科学基金资助项目(201701D22111439
201701D221121)
山西省回国留学人员科研项目(2016-088)