A joint detection algorithm of single channel GMSK mixing signals with linear complexity
HU Chang-ha · 2014
The single channel blind separation of burst Gaussian Minimum Shift Keying(GMSK) mixing signals with the same modulating parameters is studied. Based on the delay variation of mixing GMSK signals and their Inter Symbol Interference(ISI) features,a multi-branch joint decision-feedback detection algorithm is proposed. The mixing signals are taken as single signal at each branch,and the decision-feedback information is adopted to optimize the detection performance. The theory and simulation show that, the complexity of the proposed algorithm is reduced to linearity by through multi-branch decision-feedback when compared with the exponential complexity of joint maximum likelihood sequence estimation; a gain of about 2 dB can be obtained in the detection compared to that in the case of no decision-feedback; and the performance is close to that of Per-Survivor Processing(PSP) Algorithm in high Eb/N0 region. The influence of time delay on the proposed algorithm is discussed. When the amplitudes are same and the relative time delay is the integral multiple of the symbol period,the mixing signals cannot be detected. Relative theoretical derivation is performed as well. Whereas the mixing signals can be detected,the larger the relative time delay, the better the performance.