Robust quadratic time-frequency distributions for the analysis of mulicomponent FM signals in heavy-tailed noise

We consider the problem of instantaneous frequency estimation of multicomponent frequency-modulated signals corrupted by additive heavy-tailed noise. For that, a new time-frequency distribution, named the robust modified B-distribution (R-MBD), is developed as a generalization of the robust minimax M-estimates to handle such signals. We show that this representation outperforms the robust polynomial Wigner-Ville distribution (r-PWVD) in term of high resolution for this class of non-stationary signals. The proposed approach is compared to the higher-order ambiguity function (HAF) algorithm, for the instantaneous frequency estimation of a multicomponent signal. Computer simulations show the superiority of the proposed algorithm over the HAF