Track Irregularity Feature Extraction Based on the Improved Ensemble Empirical Mode Decomposition

A method of extracting transient features of a non-stationary signal was proposed based on improved Ensemble Empirical Mode Decomposition (EEMD) and the Sliding Window Singular Value Decomposition. Through fast bandpass filtering, the instantaneous frequency and amplitude can be calculated, the Intrinsic Mode Function (IMF) can be got, and at the same time, problems existing in the original EEMD method, such as mode aliasing, can be solved. The track irregularity signal is studied with the method, and the statistical and analytic result indicates that there are relatively serious short wave and medium wave track irregularities in the sample signal. Analysis of track irregularities using improved EEMD is an effective method for guaranteeing the safe of railway transformation.