Time-frequency analysis based on time-frequency-adaptive optimal-kernel

A novel method of time-frequency analysis method called time-frequency-adaptive optimal-kernel time-frequency representation (TFR) is proposed to suppress the cross-term interference caused by components at the same time. Unlike signal dependent TFR and time-adaptive optimal-kernel TFR, the time-frequency-adaptive optimal-kernel TFR obtains local ambiguity function by applying 2D windowed Fourier transform to signal's Wigner-Ville distribution, which allows the time-frequency-adaptive optimal-kernel TFR to adapt its kernel function not only to the time but also to the frequency. By applying to synthetic signal analysis, the proposed method shows a more legible result than time-adaptive optimal-kernel TFR and exhibits a better cross-term suppressing capability.