Time-frequency Analysis of Non-stationary Signal Based on Sliding Mode Singular Spectrum Analysis and Wigner-ville Distribution

Abstract

In order to obtain better performance of time-frequency representation(TFR) in non-stationary signal processing using Wigner-Ville distribution(WVD), an adaptive time-frequency analysis method based on combination of sliding mode singular spectrum analysis(SM-SSA) and Wigner-Ville distribution(SM-SSA-WVD) was proposed. Through mode decomposition of original signal by SM-SSA and linear superposition of TFRs for each mode by WVD, the cross terms are effectively suppressed, which obtaining higher adaptive ability and energy concentration of WVD results comparing to the traditional WVD methods. The proposed cross-term free method is applied to analyze several kinds of simulated clean and noisy multi-component modulated signals. The TFRs of the proposed method work better in restraining the cross terms of WVD than some other adaptive decomposition methods, thus obtaining higher time-frequency resolution and noise robustness.