Signal feature extraction based on an improved EMD method

Empirical mode decomposition (EMD) is a powerful tool for analyzing composite, nonlinear and non-stationary signals. Intrinsic mode functions (IMFs) obtained by EMD are band-limited, which can represent the features of signal and reserve the local information. Because of lacking theoretical foundation, EMD has many problems in scale mixing, stop criterion, etc. In this paper, we use an improved EMD method for signal feature extraction. Optimal envelopes mean is obtained by an inverse EMD filter scheme in the improved method. A new sifting stop criterion is proposed to guarantee the orthogonality of the sifting results. And finally, two different methods are utilized for performance evaluation of the decomposition results. Numerical simulation and experimental result demonstrate the validity of the improved method.