Single-channel blind separation of co-frequency MPSK signals based on PSP algorithm with DFSE

In this paper, the problem of blind separation of a mixture of two co-frequency MPSK signals with the same power level in a single channel is investigated. The performance of the existing per-survivor processing (PSP) algorithm which uses decision-feedback sequence estimation (DFSE) to do the detection of the two sources is firstly analyzed. It is found that the BER of the PSP algorithm with DFSE is affected by the sampling instance. A sampling instance which leads to lower bit error ratio (BER) as compared with the conventional sampling instance is then proposed. Simulation results show that by using the proposed sampling instance, the BER obtained by the PSP algorithm with DFSE is improved by about 2dB in terms of SNR.