Research of Time-frequency Analysis Method of Nonstationary Periodic Signal

Nonstationary periodic signal (NPS) is a sort of stochastic signal whose period follow normal distribution, and the statistic average value of period (SAVP) of NPS is very important but difficult to yield. Several time-frequency analysis method including short time power spectrogram (STPS), Wigner-Ville distribution (WVD) and wavelet transform (WT) are employed to obtain the SAVP of NPS, then Gauss white noise and color noise with large amplitude are used to verify the anti-interference capability of these methods. The analytic results show that STPS with inertia filtering can separate NPS from noise efficiently and has good time-frequency resolution of NPS, which means high measurement precision of SAVP. Smooth pseudo Wigner-Ville distribution (SPWVD) and redistribution SPWVD have excellent time-frequency aggregation and resolution, but their anti-interference capability is a flaw. WT has powerful ability of anti-disturbance. In sum, all the three methods can compute SAVP of NPS, among them, STPS and WVD have higher measurement precision while WT has more perfect dynamic characteristic.