A Study On Precoding Technology In Massive MIMO Systems

Massive MIMO technology can effectively improve system capacity,energy efficiency and link reliability without increasing signal transmission power,and has become one of the key technologies of 5G.The simple linear precoding algorithm can achieve optimal system performance due to the hardening characteristics of traditional massive MIMO channel.Since the system has more users,the complexity of the linear precoding inversion is relatively high.Therefore,the first research content in this thesis is to design the low complexity ZF(Zero forcing)precoding algorithm without significantly reducing the system performance.In addition,millimeter wave with rich spectrum resources has also become one of the key technologies of 5G,and the combination of millimeter wave and massive MIMO system has become one of the research hotspots in recent years.However,in millimeter wave massive MIMO systems,the hardware cost and power consumption of the digital precoding scheme in the system downlink are too high.The second research content in the thesis is hybrid precoding algorithm based on partially-connected in millimeter wave massive MIMO systems.At first,the research background and research status of precoding technology in massive MIMO systems and millimeter wave massive MIMO systems is introduced in the first chapter.And the opportunities and challenges faced by the two technologies are listed.In the second chapter,the channel model and system model of massive MIMO systems and millimeter wave massive MIMO systems are introduced in detail.The ZF and MMSE(Minimum mean square error)precoding schemes of massive MIMO systems are described deeply,and the channel capacity performances of the two schemes are simulated and compared.On the other hand,the hybrid precoding scheme based on fully-connected and partially-connected millimeter wave massive MIMO systems is also discussed in detail.And the optimal digital precoding algorithm,OMP(Orthogonal Matching Pursuit)hybrid precoding algorithm and SIC(Successive Interference Cancellation)hybrid precoding algorithm are simulated and compared.In the third chapter,the single-cell system of traditional massive MIMO system is studied,and its channel hardening characteristics are analyzed,that is,with increasing the number of transmitting antennas of the base station,the channel state information mainly distributes on the main diagonal.Aiming at the large matrix inversion problem of traditional massive MIMO systems,three precoding schemes with low complexity are proposed.The first low complexity precoding scheme is based on Lanczos method.This scheme transformed the high dimensional matrix inversion into a low dimensional three diagonal matrix inversion through Lanczos algorithm,which greatly reduces the complexity of the ZF algorithm inversion.Secondly,a low complexity precoding scheme based on weighted two diagonal matrix is proposed,firstly in which the diagonal matrix in Jacobi iteration is replaced with the lower binary diagonal matrix,and then the iterative result is weighted with the previous iteration result to obtain an iterative result of the WTM(Weighted Two diagonal Method)algorithm.In order to speed up the convergence,the first term of the Neumann expansion is taken as the initial iterative value.Thirdly,based on the Gauss-Seidel algorithm,a low-complexity ZF precoding algorithm based on weighted Gauss-Seidel algorithm is proposed.By adding the weights of the traditional Gauss-Seidel iteration result and the previous step result,the convergence speed of the algorithm can be accelerated.All the three algorithms mentioned above reduce the complexity of the inversion operation of ZF algorithms from O(K3)to O(K2),where K is the number of users,and the simulation results show that the performance loss is small.In the fourth chapter,a millimeter wave massive MIMO system with partially-connected is studied.Two hybrid precoding schemes are proposed for the hybrid precoding problem in this system.The first hybrid precoding scheme is based on band matrix that takes the achievable sum rate of the system as objective function to design the optimal hybrid precoding matrix by pre-designing digital precoding matrix as a band matrix.The second scheme obtains the optimal hybrid precoding matrix by getting the minimum difference between optimal digital precoding matrix and hybrid precoding matrix.In order to reduce the system hardware cost and power consumption,3-bit quantized phase shifters is proposed in the thesis proposed to obtain optimal hybrid precoding matrix.The simulation results show that the two proposed algorithm has certain advantages in performance and complexity.At the end,the summary and prospect are given,and the shortcomings of the paper are summarized and some issues are open.