The Research And Implement Of Adaptive Multiuser Detection In 3G

In 3G, MAI (Multi Access Interference) may reduce system capability and performance seriously. In order to eliminate its influence essentially and improve system performance, we must use multiuser detect technology. Multiuser detection (MUD) is one of the key technologies in 3G mobile communication system, and has been a hot spot in telecom research field. In recently years, the main research work may focus on how to present a high performance, low complexity algorithm.This paper summarize the recent multiuser detect algorithm, and study blind adaptive multiuser detection. The main work is showed as follows:Based on the study of CMA (Constant Module Algorithm), an Update Constant Module Algorithm (UCMA) is proposed for time-varying channel, and a new performance parameter-MVC (module varying cost) is presented, it is designed to figure the compute cost in time-varying channel. Computer simulations result shows that the algorithm can speed up the convergence rate.The second work is applied the UCMA into Space-Time two-dimension processing. S-T multiuser detection utilizes the information both of spatial and time diversity, so can reduce MAI and ISI together. Simulate results show that it will gain better SIR than only in time processing.In order to settle the question of bad system performance in low SNR environment, this paper present a improved constrained optimization multiuser detect algorithm. This algorithm joint the constrained optimization algorithm and subspace algorithm, can improve the system performance in strong interfere environment. Simulation shows that the proposed multiuser detect algorithm can improve output SIR in low SNR Channel.In order to make the study is not only bound on theory scope, this thesis introduces a method of implementing CMA multiuser detect algorithm using FPGA(Field Programable Gate Array). FPGA is a kind of ideal device to implement MUD, because it can be reconfigurated by software, and can construct parallel unit. The method has taked both hardware resource and calculation quantities into account, so it is a practically one.