Research Of Blind Multi-user Detection For DS-CDMA System Based On Reduced-Rank Subspace

Multi-user detection is an efficient interference suppression technique for 3G that can reduce the multi-access interference (MAI) and improve system capacity. Recently, blind multi-user detection without requiring information of interference users has attracted considerable attention and interests.Firstly, the basic theory, the system model and the performance measure of the multi-user detection are introduced.Secondly, the blind multi-user detection based on subspace is discussed. And then, on the basis of subspace tracking, dimension estimation, rank reducing and the dynamic condition processing, blind multi-user detection based on reduced-rank subspace is discussed and researched.Thirdly, a new reduced-rank subspace selection algorithm via variable threshold is proposed. By using the new algorithm, a proper reduced-rank subspace is obtained quickly and the result can be reused. The improved Akaike information criterion (AIC) is adopted to estimate the dimension of the signal subspace during the subspace tracking. The criterion can reduce the computation of dimension estimation on the basis of the same error probability. When the dimension of signal subspace is overestimated, the performance of the blind reduced-rank multi-user detector is analyzed.Finally, to deal with the situation that the dimension of the signal subspace decreases, a new algorithm is proposed based on the idea of the Cross Spectral Method. The simulation results indicate that the blind adaptive multi-user detection algorithm proposed in the paper shows superior detection performance.