Research On Hybrid Precoding Algorithm Based On Millimeter Wave Massive MIMO

With the rapid development of mobile Internet technology,wireless data services are also showing explosive growth,and it will be difficult for traditional frequency bands to meet people’s growing demands.The combination of millimeter wave(mm Wave)and massive multiple-input multipleoutput(MIMO)technologies have propelled the development of The Fifth-Generation Mobile Communication System(5G).Compared with traditional low-band communication,the available bandwidth of the mm Wave band is larger,which can significantly increase the system capacity.In addition,because of its short wavelength,massive antenna arrays can be integrated into a small physical space and provide sufficient array gain to reduce the free-space path loss of mm Wave signals,while attenuating some noise interference in communications,improving communication quality.In full digital precoding,each antenna is connected to a radio frequency(RF)chain,which has high cost and power consumption.In order to solve this problem,this paper aims at the design of hybrid precoding for single-user mm Wave massive MIMO system.A series of studies have been carried out on the problem,mainly based on the fully-connected(FC)structure and the sub-connected(SC)structure,and comprehensively analyzed the spectral efficiency,energy efficiency and complexity of the hybrid precoding algorithm.The main research contents are as follows:1.Chapter 1 explains the background and significance of the hybrid precoding algorithm in mm Wave massive MIMO systems,introduces the research status of the two hybrid precoding schemes at home and abroad,and explains the research content and structure of this paper.2.Chapter 2 firstly introduces mm Wave technology based on mm Wave characteristics and channel model;then introduces massive MIMO system from two aspects of massive MIMO technology and system model;finally,the precoding technology is summarized from three aspects:digital precoding,analog precoding and hybrid precoding.3.Chapter 3 presents a low-complexity hybrid precoding algorithm based on FC structure.Using an iterative method,the analog precoding matrix and the digital precoding matrix are found to be the closest to the all-digital precoder.Compared with the Orthogonal Matching Pursuit(OMP)algorithm,the algorithm proposed in this chapter does not require any assumptions about the geometry or other constraints of the antenna array,and is less complex to implement.The simulation results show that when the signal-to-noise ratio and the number of data streams are increased,the performance of the proposed algorithm is close to that of the all-digital precoding algorithm,and is significantly better than that of the OMP algorithm.4.Chapter 4 presents a new hybrid precoding algorithm based on SC structure.First,because all RF chains are connected to a set of antennas,based on this structure,a modular diagonal array can be constructed;secondly,with the goal of maximizing the total spectral efficiency,the spectral efficiencies of all sub-arrays are optimized one by one,and find The analog precoding matrix;finally,the digital precoding matrix is obtained by means of the least squares(LS)method.By optimizing the analog precoder and digital precoder separately,instead of the joint optimization problem,each optimization problem eliminates the complex iterative process.Simulation results show that the algorithm can achieve a compromise between energy efficiency,spectral efficiency and complexity.5.Chapter 5 proposes a SC structure hybrid precoding algorithm based on water-filling algorithm.In the algorithm in Chapter 4,when the number of data streams is greater than 1,only the same number of RF chains can be used to obtain better algorithm performance.In order to solve the limitation between Ns and RF chain,this chapter first uses the method in Chapter 4 to find the analog precoder;then,based on the water-filling algorithm,the power distribution of Ns is performed to solve the digital precoder.The simulation results show that when the number of RF chain at both ends of the system is greater than the number of Ns,the proposed algorithm can have higher spectral efficiency than the algorithm proposed in Chapter 4,and is close to the OMP algorithm.6.Chapter 6 summarizes the main research contents and results of this paper,points out the shortcomings,and looks forward to the next research focus.