Blind Estimation of Wide-band Frequency-Hopping Signal Based On Compressed Numeral Characteristics

In order to make a precise and quick estimation of the wide-band Frequency-Hopping(FH) signal, a blind parameter estimation using numeral characteristics of compressive samplings is proposed. First study the compressed numeral characteristics (CNC) of the signal on the premise of non-reconstructing the original signal to detect the changes of signal frequency. Then reconstruct these data sections to get its frequency domain coefficients. Estimations could be achieved with hopping frequency, frequency-hopping periods and the frequency-hopping moments based on the frequency domain coefficients. Simulation results show that this algorithm can achieve the same error conditions as Wigner algorithm with Gaussian white noise when SNR is 10dB and compress rate is 1/4. And the number of sampling points of processing required for a quarter of the Wigner algorithm. So this algorithm has the advantages of low computing complexity and better real-time performance compared with then traditional algorithm using time-frequency characteristics.