Research On Detection Algorithm Of Uplink Signal In Massive MIMO System

Massive Multiple-input Multiple-output(MIMO)is one of the key technologies of 5G,which can improve the spectrum utilization and energy efficiency by deploying a large number of antennas at the base station to serve multiple users simultaneously.However,with the increase of the number of users served by the base station,the receiving signal will become more complex and higher requirements are imposed on the signal detection algorithm.Through the study of linear and nonlinear signal detection algorithms for Massive MIMO systems,this thesis focuses on the analysis and improvement of the message passing algorithm and the approximate message passing algorithm in the nonlinear detection algorithms.The main research contents of this thesis are:(1)The message passing algorithm is based on a probability graph model,in which the Gaussian tree approximation(GTA)algorithm is suitable for the high order modulation of the Massive MIMO systems.Aiming at the serious interference of multiuser communication in the Massive MIMO systems,this thesis combine with the idea of serial interference cancellation to improve the GTA algorithm.Firstly,the improved GTA algorithm uses the Gaussian tree distribution to approximation the posterior distribution and solves the convergence problem of the message in the loop graph.Secondly,the column norm sorting method is adopted to reduce the error propagation in serial interference cancellation,which further improves the detection performance.Through simulation and verification,the improved GTA algorithm can significantly improve the detection performance when the number of users is similar to the number of base station antennas.(2)The approximate message passing algorithm has the advantages of simple structure and low complexity.In order to further improve the detection performance,this thesis combine the approximate message passing base on large system(AMP-LS)algorithm and a branch-optimized MIMO detection technique,then the algorithm base on Branch optimized MIMO detector(B-AMP-LS)algorithm is proposed.Firstly,the B-AMP-LS algorithm make Gaussian approximation of the transmitted message,and the mean and variance of the message are passed between the nodes in the system factor graph.Secondly,the B-AMP-LS algorithm introduces the compensation information for the prior information in order to calculate the external information which the decoder needs,this method improves detection performance.When the number of users is similar to the number of base station antennas,the simulation results show that the detection performance of the B-AMP-LS algorithm is obviously better than that of the AMP-LS algorithm,while both the B-AMP-LS algorithm and the AMP-LS algorithm have lower complexity.