间歇信号的经验模态筛选方法

he empirical mode decomposition (EMD) is a novel decomposition method for analyzing nonlinear and non-stationary signals. It has attracted more and more attentions in 1-D/2-D signal processing in resent years, but the frequent occurrence of mode mixing has limited its application fields. As the natural events always happen intermittently, the empirical modes to describe the events also need to be intermittent. But the definition of intrinsic mode function (IMF), which the EMD is based on, puts little emphasis on the intermittence of mode. The overall objective in this study is to avoid the frequent occurrence of mode mixing by presenting a novel conception of intermittent empirical mode function (IEMF) to describe the intermittent events and a novel sifting algorithm to sift the IEMF from original signal. Furthermore, An empirical method for periodic spectrum analysis is presented to judge whether the mode mixing occurs between IEMFs or between IMFs. Experiments on simulated signals from pure modes and the modes corrupted by Gaussian noise are made with a view to demonstrating the validity of presented methods. The results of the experiments indicate that the proposed sifting algorithm can overcome Mode mixing, and the physical significance of the IEMF is more definitude