Abstract:

In view of the problem that conventional nonlinear timefrequency analysis methods in frequency hopping(FH)signals parameter estimation suffer from the effect of serious crosscomponents, a radially Gaussian kernel(RGK)timefrequency analysis method is introduced. To suppress the crosscomponents, it selects the adaptive optimal kernel depending on a variety of signals. The RGK timefrequency analysis method can estimate FH signals parameters in Gaussian noise, but its performance in heavytailed impulsive noise environment falls into severe degradation. Combined with the maximum likelihood estimation theory, a weighted maximumlikelihood generalized Cauchy(WMGC)method for the case of α stable distribution noise is proposed. The parameters of noisy FH signals can be estimated by the RGK timefrequency analysis method based on the WMGC filter(WMGCRGK method, simply WR method). Simulation results show that compared with the fractional lower order statistics as well as the Myriad filter based time frequency analysis methods, the proposed method has better performance on the FH signals parameter estimation and it is robust to the α stable distribution noise.