Adaptive spectrogram vs. adaptive pseudo-Wigner-Ville distribution for instantaneous frequency estimation

An approach that automatically adapts the window length to the changes in instantaneous frequency (IF) and tracks it better than a fixed window approach is used to estimate the IF using the peak of the spectrogram. An adaptive window-based time-frequency representation is more useful for tracking events in time and frequency. The peak of the spectrogram obtained using the adaptive window length algorithm is used as an IF estimator and its performance in the presence of multiplicative and additive noise is studied. The performance is compared with that of the pseudo-Wigner-Ville distribution (Ps.WVD). Both analytically and experimentally, the adaptive spectrogram is found to be more robust than adaptive Ps.WVD