A Modified Empirical Mode Decomposition Method with Applications to Signal De-noising

Abstract

In order to solve the problem of nonlinear and nonstationary signal de-noising, a novel mode cell filtering(MCF) method was proposed based on empirical mode decomposition(EMD).The method defined the signal between the two adjacent zero-crossings within intrinsic mode function(IMF) of EMD as a mode cell, and treated the mode cell as the basic analyzable object.The mode cells were sorted by judging the amplitudes of the mode cells, and then the mode cell s filtering model was established.Evolutional rules of the amplitude of the mode cell were analyzed when the noisy signal corrupted by fractional Gaussian noise with different Hurst exponent were decomposed by EMD method, and threshold choosing rules used in Gaussian de-noising were also established.Numerical simulation and real data test were carried out to evaluate the performance of the method.Simulation and test results showed that the proposed method outperformed the optimal wavelet threshold de-noising algorithm in whole, so it is a novel effective signal de-noising method with virtue of self-adaption.