Research And Design Of Multi-user Detection Algorithms For Direct Sequence Spread Spectrum CDMA System

Multi-user detection technology is an important means to solve the problem of multiple access interference and near-far effect in CDMA communication system.In this paper,we mainly study the design of robust multiuser detector based on new intelligent computing method,and blind estimation method of PN sequence in DS-CDMA system under complex channel environment.The innovation and main works of this paper are as follows:(1)Traditional multi-user detection algorithms are work on ideal Gaussian channel,but there are a lot of non-Gaussian noises exist in the actual channel environment,such as impulse interference and so on.In this paper,? stable distribution model is used to model the impulse noise channel,and a novel robust multiuser detector is constructed by using Huber penalty function of M estimator.(2)In this paper,a hybrid grey wolf optimization algorithm is proposed based on the combination of the improved grey wolf optimization algorithm and JADE algorithm by the strategy of evolutionary direction information,and the proposed algorithm are applied to solve the optimal multi-user detection problem.Firstly,the evolutionary direction vector information of the three wolves in grey wolf optimization is used to update the location of the individual in order to replace the original arithmetic average position updating method.Then the fitness of three wolves is compared with the fitness of new individuals generated by the crossover operation of JADE algorithm.If the fitness of the new individual is better than any one of the three wolf individuals,the new evolutionary direction vector is obtained by subtraction between the location information of the corresponding wolf species and the new individual.The new evolutionary direction vector is preserved in the archives of JADE algorithm in that to enhance the reliability and robustness of the hybrid algorithm optimization process.To further illustrate the performance of the new hybrid algorithm,we introduce several other new intelligent optimization algorithms as a performance comparison.(3)Aiming at non-cooperative communication,an anti-impulse noise blind estimation method of pseudo-random sequence of direct sequence signals based on de-noising preprocessing matrix using fractional lower-order statistics and M estimation are proposed in this paper.The new method can effectively suppress the impulse component in the received signal and reduce the adverse impact of impulse interference on the sequence estimation algorithm.Meanwhile,a novel method for estimating the out-of-step point of pseudo-random code sequence is proposed by using the information of signal-to-noise amplitude difference.(4)In this paper,we mainly study the blind estimation of PN code sequence based on matrix decomposition,and verifies and systematically summarizes the related means under ideal Gaussian channel through simulation.In addition,the PN code estimation performance of the associated algorithm under impulse noise channel is tested in detail by using the proposed de-noising preprocessing method.