Research On Low-complexity Linear Iterative Precoding Algorithm In Large-scale MIMO Systems

The large-scale MIMO systems have been recognized as one of the key technologies for future mobile communications,i.e.,configuring a large number of antennas at the base station and simultaneously serving dozens of user terminals,which makes the spatial freedom more.Moreover,it can also achieve huge array gain and multiplexing gain.It has been theoretically proved that the spectral efficiency can be increased by more than two orders of magnitude.Since the increase of the user terminals may cause interference among users,nonlinear or linear precoding must be used to suppress interference before the base station transmits data.However,the traditional linear ZF precoding involves channel matrix inversion.Therefore,its computational complexity increases as the size of the antenna array increases.In terms of the above problems,the low complexity linear precoding scheme is studied in this paper.In this paper,the computational complexity of the channel matrix inversion in linear ZF precoding is reduced by using the iterative idea.Firstly,an improved weighted two-stage iterative(WTS)precoding method is proposed against the shortcomings of the existing iterative methods with less iterations and slower convergence rate,which combines the forward and backward iteration results by the empirical weighting parameter so that adjusting the spectral radius of the iterative matrix further affects the convergence rate.Secondly,in view of the above shortcomings,and considering the effect of relaxation factor and iterative initial solution on convergence rate,an improved modified successive overrelaxation(MSOR)precoding method is proposed,i.e.,through reasonable selection of relaxation factors and acceleration factors to speed up convergence.Compared with the traditional linear ZF precoding,the proposed two improved precoding methods have lower computational complexity.And the convergence rate of the two improved precoding methods is better than the existing precoding methods.The simulation results show that the proposed two improved precoding methods can obtain the approximate optimal linear ZF precoding performance with fewer iterations.Therefore,the proposed precoding methods are more suitable for large-scale MIMO systems.