Study On Blind Multiuser Detection In CDMA System

Multiuser detection (MUD) is one of the key technologies in 3G mobile communication system, and has been a hot spot in telecom research field. MAI (Multiple Access Interference) may reduce system capability and performance seriously in 3G. In order to eliminate its influence essentially and improve system performance, we must use multiuser detection technology. In recently years, the main research work may focus on how to present a high performance, low complexity algorithm. Traditional multiuser detector make good use of all signals information which result in multiple access interference, so that it provides optimum MAI resistance. But the frequent use of training sequence is certainly a waste of channel bandwidth.Research shows that with the prior knowledge of only the signature waveform and timing of the user of interest, blind adaptive multiuser detector can effectively detect data symbol of the desired user without training data sequence for every active user. It is especially attractive for the downlinks system and suppressing intercell MAI effectively.This paper based on much blind multiuser detection technologies. Firstly, Multiuser detection in CDMA system is researched, then two modified Constant Modulus Algorithms (CMA) are proposed: variable step size CMA and modified linearly constrained CMA. Also multiuser detection based Kalman filter is researched, the LMS algorithm and RLS algorithm are studied by computer simulating. At last, the simulating conclusion show that the two new CMA have much more capabilities than the old ones, and have low complexity. Whether in static or dynamic situation, the Kalman filter algorithm has a good capability of convergence and tracking in AWGN channel, also the simulation results show that it has strong capability of suppressing MAI.