Abstract:

A parameter blind estimation algorithm of Costas frequency hopping signals is presented. The fast Fourier transform (FFT) is used to estimate the number of hopping frequencies by frequency spectrum analysis, then the initial estimation of code width is obtained. The accurate estimation of code width can be effectively extracted by searching the frequency hopping position with the search coverage set according to the initial estimation. In the condition of unknown pulse width, a time of arrival estimation algorithm based on minimum description length (MDL) is presented. Compared with other algorithms, this algorithm has the characteristics of shorter computation time and higher parameter estimation accuracy. Simulation results indicate that it is effective even at a low signal to noise ratio (SNR).