Multi-user Detection In Cdma Applications

Multiuser detection is one of the key techniques in the mobile communication system .The main idea is that it detects the signal customer based on all signals, which results in MAI. Thus it has good anti-interference function, relaxes the 'near-far' effect, and increases the capacity of system. In this paper the multiuser is studiedFirstly, several kinds of typical multiuser are introduced, but those algorithms acknowledge the training sequences. Merits and shortcomings are analyzed for those algorithms .And those adaptive algorithms are also analyzed. The blind multiuser is extensively concerned for it has no needs of the training sequences. Secondly three kinds of blind adaptive algorithms with direct methods ,subspace methods and CMA are discussed respectively by the theoretical analysis and the simulate results.In this paper a new adaptive step-size is proposed on the MOE algorithm. It make full use of the projection matrix and LMS algorithm . Compared the fixed step-size and adaptive step-size of LMS algorithm, simulation results show that it has the character of convergence fastly , SIR highly and BER lowly . In this paper .The convergence of LCCMA algorithm is also mainly analyzed in this paper. Simulate results show that the step-size is related with .convergence .Finally the RAKE receiver and subspace are applied in the frequency-selective slowly multi-path channel to estimate the coefficient of channel.