Comparative performance analysis of two algorithms for instantaneous frequency estimation

This paper compares two algorithms for estimating the instantaneous frequency of complex signals: the high-order ambiguity function (HAF) and the polynomial Wigner-Ville distribution (PWVD). Comparison is made by asymptotic first-order error analysis, which is verified by simulation. It is shown that when the signal phase is a polynomial function of time, the HAF always outperforms the PWVD. Two other advantages of the HAF over the PWVD are: (a) it has lower computational complexity; (b) unlike the PWVD its use is not limited to the instantaneous frequency at the middle of the observation interval.