Research On Precoding And Power Allocation Of Cell-Free Massive MIMO Downlink

The fifth generation mobile communication network(5G)has been deployed on a large scale in China,and more and more users use the 5G mobile network with high data rate and low latency.The research on the beyond 5G and 6G is also on the road.Cell free large-scale MIMO(Multiple Input Multiple Output)is considered as a promising technology.It has broken through the traditional cellular architecture and changed the network architecture from cell central to user central.The network no longer has cell boundaries.Through the cooperation between APs,the inter-cell interference can be significantly eliminated,path loss can be reduced,and system performance can be improved.The research of this thesis mainly focuses on the downlink precoding and power allocation of this system.This thesis first introduces the cell free massive MIMO system and its related theories,as well as the principle of Rate Splitting Multi Access(RSMA)access technology,and compares the differences and advantages between RSMA and Space Division Multiple Access(SDMA),nonorthogonal multi access(NOMA).Then this thesis studies the downlink precoding and power allocation of radio strips under SDMA and RSMA access respectively.The main contributions are as follows:Firstly,a distribute cooperative zero forcing downlink precoding and power allocation scheme is proposed for the radio stripes,which is constrained by the limited backhaul and the error generated by the centralized processing.At the same time,inspired by the fully distributed regularization zero forcing precoding,a practical regularization factor design scheme is given.The simulation results show that the two precoding schemes have different performance improvements compared with the conjugate beamforming precoding without pilot contamination,but they are still inferior to the centralized zero forcing precoding with unlimited backhaul capacity.The performance of local cooperative zero forcing precoding is degraded or even worse than that of conjugate beamforming precoding with pilot contamination.Secondly,this thesis discusses the system design when using RSMA access in the large-amount users access scenario,and proposes a low complexity downlink public and private information precoding scheme for RSMA Cell Free Massive MIMO system using the linear combination of full rank matrix of channel estimation.At the same time,a power allocation algorithm based on the combination of weighted minimum mean square error and alternating direction method of multipliers to maximize the system throughput is proposed to replace the successive convex approximation algorithm.The simulation results show that RSMA can reduce the influence of pilot pollution under specific access point-user distributions,and the proposed algorithm greatly reduces the large computation time and little performance compared with the successive convex approximation algorithm.Finally,it summarizes the thesis,and points out the shortcomings of this thesis and the research directions and problems to be solved in the future.