Abstract: A method of single channel source separation and parameters estimation of multi-component PRBC-LFM signal is proposed. The repetition period of each component can be estimated through singular value decomposition. The influence of the pseudo code on parameter extraction is eliminated through a square calculation of signal, and the interference of noise on the component under separation is reduced by cycle accumulating. The modulation rate and the carrier frequency are estimated by using FRFT and searching the discrete points. The initial phase of multi-component signal is determined by calculating the inner product, and then the pseudo random sequence and the signal amplitude are estimated by using the character of that the pseudo random sequence is a real signal. In addition, the proposed method is used to estimate SNR of mixed signal by using eigenvalue decomposition, and SNR can be used to determine the threshold adaptively. In simulation, the method is demonstrated by separating a specific signal. The result shows that the proposed method is effective under the condition of different SNRs.

Key words: ordnance science and technology, multi-component signal, PRBC-LFM signal, parameter extraction, source separation, adaptive threshold