Track Irregularity Analysis Based on Empirical Mode Decomposition

The principle and the course of signal filtering and denoising using empirical mode decomposition (EMD) are described, and simulative signals with wavelet and EMD are compared, and the filtering, denoising method of EMD is introduced to process the signals in track irregularity analysis field. The results show that both methods of wavelet and EMD are effective, while the method based on EMD, which filters signals and eliminates the noise selectively according to spectrum character of decomposition signal, is the optimal method that can sufficiently keep the non-stationary character inhered in signal and provide well auto adaptability, flexibility and affectivity and feasibility in filtering and de-noising of track irregularity signals.