Research On Hybrid Precoding Technology For Millimeter Wave Massive MIMO Systems

The explosive growth of mobile data services and intelligent terminal applications has put higher requirements on the coverage,transmission rate and energy consumption of wireless communication networks.As one of the enabling technologies of future mobile communication,millimeter wave massive multiple input multiple output(MIMO)technology has significant advantages such as large bandwidth,high-frequency spectral efficiency and high reliability.In the millimeter wave massive MIMO systems,the hybrid precoder can overcome the free-space path loss and signal interference of millimeter-wave signals,thereby establishing a high signalto-noise ratio transmission link.Therefore,design of the hybrid precoder in the millimeter wave massive MIMO systems to achieve the balance between performance has become a hot issue to be studied.The contributions of the thesis are summarized as follows:(1)Aiming at the loss of spectral efficiency caused by the limited resolution of phase shifter,a hybrid precoding scheme using an off-grid quantized analog precoding codebook is proposed.The Taylor series expansion is used to form the initial codeword of the column vector of the analog precoder.According to the orthogonality correlation between the orthogonal projection of the optimal unconstrained precoding matrix and the codeword,the unary quadratic function with the argument of the quantization angle error is used to update the angle quantization grid,approximate the optimal quantization angle,and improve the quantization accuracy to update the analog precoder.The experimental results show that in the case of low resolution phase shifters,the proposed algorithm can significantly improve the spectral efficiency and reduce the requirements for phase shifters.(2)To improve the spectral efficiency and reduce bit error rate,a joint sparse hybrid precoding optimization algorithm is proposed.Firstly,the millimeter wave spatial lobe channel is reconstructed,and the hybrid precoding optimization design is carried out for each lobe subchannel respectively.Then,aiming at the problem of multivariate sparse signal reconstruction with non-convex constraints,the self-support set and common support set of each data stream analog precoding in the lobe are designed by using the implicit sparsity structure of the digital precoder to jointly optimize the analog precoding matrix and digital precoding matrix.Simulation results show that the proposed algorithm can achieve higher spectral efficiency and better bit error rate performance.(3)In order to explore a more flexible hybrid precoding scheme,a hybrid precoding structure based on dynamic-partially connection is proposed.By adjusting the switching state between RF chain and antenna,the number of phase shifters is adjusted,and a phase alternating optimization design is proposed to optimize the phase of phase shifters column by column,so as to improve the spectral efficiency performance.And given the target spectral efficiency,a dynamic index updating algorithm is designed.Taking the antenna selection strategy into account,the hybrid precoding connection structure is designed by maximizing the gain of largescale antenna array.The proposed hybrid precoding method reduces the number of phase shifters required by jointly optimizing digital precoding,phase shifter network and switch connection network,and reduces the hardware power consumption while improving the spectral efficiency.