Time-varying Parameter-based Synchrosqueezing Wavelet Transform with the Approximation of Cubic Phase Functions

Synchrosqueezing wavelet transform (SST) is a linear time-frequency analysis method, which aims to sharpen the distribution of the continuous wavelet transform (CWT) and separate multicomponent signals. In this paper, we propose a new SST algorithm with the time-varying parameters and second-order phase transform. More precisely, we analyze the separation abilities and conditions of CWT and SST by approximating a non-stationary signal with cubic phase functions. The optimal time-varying parameters are also proposed. The experimental results show the validity and efficiency of the proposed method, especially for multicomponent signals with fast varying frequencies.