The Research Of Multi-user Detection Technologies For CDMA Communication Systems

Code Division Multiple Acess (CDMA) is a self-interfered system, where multiple access interference (MAI) and near-far effect existing as the main factors restricting CDMA system capacity. Multi-user detection is one of the key techniques in the 3G mobile communication systems, which eliminate the multiple access interference and relaxes the near-far effect, and consequently increase the capacity of the system.This thesis focuses on the recent multi-user detection algorithms, and an improved multi-user detection algorithm based on the conventional constant modulus algorithm is given. The conventional multi-user detection methods which are under the condition of AWGN (additive white Gaussian noise) channel assumption will incur significant performance degradation when there is impulsive noise in the channel. However, many physical channels, in the real situation, are not AWGN but coexisted with the impulsive noise. An important non-Gaussain distribution model calledαstable distribution is introduced. Here, theαstable distribution is adopted to imitate the impulsive noise occurring in many physical channels, and several new blind multi-user detection methods for impulsive noise assumption are been proposed. The main research work and contributions of the thesis are summarized as follows:①The blind multi-user detection method based on constant modulus algorithm (CMA) is further researched. The convergency performance of CMA is affected by the iterative step. To solve this problem, a modified method called optimal interative step method is proposed. The simulation results indicate that the modified method offers a faster convergency speed.②The main research focus on the blind multi-user detection algorithm in impulsive noise. Theαstable distribution is adopted to imitate the impulsive noise channel. Based on the theory of generalized normal-distribution signal, several new blind multi-user detection methods for impulsive noise channel assumption are proposed. These methods are based on least p-norm, generalized CMA and least squares generalized CMA. Simulation results indicate these methods are robust and can resist MAI in both AWGN channel and impulsive noise channel.