Research On 3D Large Scale MIMO System Precoding Technology

With the rapid development of communication technology,5G era has begun to come.The traditional multi-input multi-output(MIMO)system can no longer meet the daily needs.3D MIMO can expand the vertical dimension channel information by configuring the two-dimensional active antenna array at the base station end and provide abundant spatial freedom in horizontal and vertical directions.Therefore,it has been widely studied by domestic and foreign researchers.Among them,precoding technology,as one of the core technologies of 3D MIMO,can effectively reduce the interference between receiving users and improve the performance of 3D MIMO system when the base station gets the downlink channel information.However,in the FDD system,the base station can only obtain the downlink channel information by means of up-link feedback,which leads to too much overhead and too long feedback time due to too much instantaneous channel information.Thus,the channel characteristics cannot be fully reflected.In view of this problem,this thesis considers the use of channel statistics,channel information and JSDM technology to study 3D MIMO precoding and other key technologies.First of all,this thesis introduces in detail the process of extending the Two Dimensional Multi Input Multi Output(2D MIMO)system to the Three Dimensional Multi Input Multi Output(3D MIMO)system,and explains the parameters in the 3D MIMO system in detail.At the same time,the process of channel modeling from 2D MIMO to 3D MIMO is analyzed.Then several classical 3D MIMO precoding algorithms are derived in detail and the advantages and disadvantages of each algorithm are compared.On this basis,a hierarchical precoding algorithm based on region division is proposed.This algorithm firstly divides 3D MIMO precoding into horizontal dimension and vertical dimension precoding by using the properties of klonecker product,and then carries out interference elimination in vertical dimension.Since the hierarchical precoding only needs the statistical state information of the channel instead of the instantaneous channel state information,it can reduce the computational complexity.Secondly,in order to reduce the error rate of 3D MIMO system and improve the system throughput,this thesis give the other a low complexity of two levels of precoding algorithm based on dimension reduction after the user channel of transfer matrix.The algorithm first groups the users,and then precodes the users.According to the Kronecker product structure of 3D MIMO channel,the first layer of precoding is divided into horizontal dimension and vertical dimension.For the interference of users in the group,the second layer of precoding adopts the zero forcing criterion.Because the layered precoding only needs the statistical state information of the channel and does not need the instantaneous state information of the channel,it can reduce the pilot training overhead.Finally,in order to improve the system capacity of 3D MIMO,another low complexity two-level precoding algorithm is proposed based on the user channel transmission matrix after grouping.Based on the maximum SLNR criterion,the first level precoding is designed to suppress inter group interference.The second level precoding is designed to eliminate intra group interference,which can be processed by ZF precoding.Simulation results show that the proposed algorithm can achieve better performance than BD-ZF precoding scheme.