Research On Hybrid Precoding In Millimeter Wave Massive MIMO Systems

One of the target of the 5th generation(5G)mobile communication systems is to inprove the transmission rate.Millimeter wave has abundant spectrum resources,which can increase bandwidth.Antenna arrays deployed by massive multiple input multiple output(MIMO)can improve spectrum utilization.Therefore,millimeter wave and massive MIMO occupy a very important position in 5G.Precoding is one of the key technologies of the MIMO system.Traditional analog precoding is simple to implement,but only supports single data stream transmission and has poor performance.Digital precoding can be extended to multiple data streams,but it limits the number of radio frequency(RF)chains.Hybrid precoding combines the advantages of analog precoding and digital precoding.And it can be well applied to massive MIMO scenarios.This paper focuses on the research on the millimeter wave massive MIMO hybrid precoding algorithm.The main work is as follows.Firstly,the propagation characteristics of millimeter waves are studied,and a channel model suitable for millimeter wave communication is introduced.Then,the spectrum efficiency of single-user and multi-user massive MIMO systems are deduced respectively.Next,the classic MIMO system precoding algorithms are introduced.In the single user scenario,aiming at the problems of loss channel capacity and high algorithm complexity of the existing matrix decomposition-based hybrid precoding algorithms,a two-stage based low complexity hybrid precoding algorithm is proposed.The algorithm is divided into two parts,obtaining the optimal full-digital precoder and the hybrid precoder.Firstly,this paper combines singular value decomposition(SVD)and water filling algorithm to design an optimal full-digital precoder with the requirement of lossless capacity.Then,a hybrid precoding matrix is designed with the goal of maximizing spectrum utilization.To reduce the complexity of searching columns of the overcomplete matrix,the two-stage hybrid precoding(TS-HP)algorithm is proposed to obtain the hybrid precoding matrix.Complexity analysis and simulation results show the proposed algorithm has lower complexity than traditional precoding algorithms,and has better bit error ratio(BER)and spectrum utilization.In the multiple user scenario,aiming at the existing signal-to-leakage-plus-noise ratio(SLNR)-based hybrid precoding algorithm that do not take into account the problems of different user channel quality and ill-conditioned channel matrix,a dual-layer hybrid precoding algorithm that combines SLNR and vector perturbation(VP)is proposed.The algorithm designs the precoding matrix in two layers.In the first layer,considering the different user channel quality,a matching weighting algorithm based on the SLNR criterion is proposed.The algorithm consider the influence of channel coefficients,precoding vectors and combination vectors simultaneously to design channel weighting factors.The weighting factors are used to improve the SLNR algorithm and design a better hybrid precoder to eliminate inter-user interference.In the second layer,according to the equivalent baseband channel obtained from the first layer,zero force(ZF)is used to eliminate the interference between antennas.Considering the ill-conditioned channel matrix,VP algorithm is used for further processing.The perturbation vector is added to the transmit signal to compensate the influence of the increase of transmit power.Complexity analysis and simulation results show the proposed algorithm has better BER and spectrum utilization than similar hybrid precoding algorithms.