Research On Recovery And Sorting Of FH Signals Based On UBSS

Because of the good anti-jamming capability and low probability of interception, frequency hopping(FH) signal is widely used in military applications. It has great reality significance to carry research on frequency hopping investigative technique. In order to obtain the time domain data of FH signal, we, the recovery of underdetermined mixing FH signal is studied from reconstructing signal each hop based on underdetermined blind signal separation(UBSS) and FH signal sorting, under the assumption that the frequency hopping signal parameters is known.Firstly, aiming at solving the problem that the method by time-frequency domain conversion has a high complexity and can’t get the time domain data of FH signal, an algorithm based on filtering to meet the sparsity requirements is proposed. With filtering, the UBSS of FH signal is divided, and every component obtained by dividing meets the requirements of sparsity for UBSS. Then with UBSS of every component the time domain data of each hop signal can be got. The simulation results show that the algorithm reduce the complexity, while the the accuracy of the recovery is high. Compared with the method by time-frequency domain conversion, it is more efficient for the UBSS of FH signals.Secondly, the sorting technology for synchronous orthogonal frequency-hopping signal is discussed to remove the vagueness brought by UBSS. By extarcting the information of direction of arrival(DOA) from mixing matrix and estimating the symbol rate, a sorting algorithm based on multi-parameter fusion is proposed. The simulation results show that the algorithm has a good performance in terms of sorting frequency-hopping signal. When the signal to noise ratio(SNR) is higher than 8 d B, the sorting probability can reach more than 0.9.