Abstract：Because short term fourier transform did not have high time-frequency differentiability, wavelet transformation had too much operation time and Wigner-Ville distribution ( including its improved methods) had cross-term interference. Thusly, a method called Hilbert-Huang Transform ( HHT) was put up with to estimate the parameters of frequency-hopping signals. Its decomposition was adaptive and had high time differentiability and frequency differentiability, but it had end effects and illusive components. The end effects can be resolved by using a mirror closed extending method, and illusive components can be treated by using a correlation coefficient. Simulation results showed that this improved algorithm could solve the problems effectively.

Keyword：parameter estimation;frequency-hopping signal;HHT;EMD;