Optimal kernel design and time-frequency analysis for frequency hopping signal using entropy measure

In this paper we have proposed an optimal procedure of bi-directional Gaussian kernel based on third order entropy measure with volume normalized and applied to estimate hop timing, hopping frequency and hop during of FH signals embedded in additive white Gaussian noise. The estimation performance has been evaluated by comparison with signal-dependent radially Gaussian kernel. Simulation results have shown that the proposed method presents modest improvements in the threshold SNR and estimation variances over the latter.