Research On Channel Estimation And Precoding Optimization Algorithm Of Mm Wave Massive MIMO System

Compared with the fourth generation of mobile communication system(4G),the fifth generation of mobile communication system(5G)has greatly improved in the peak rate,spectrum efficiency and regional communication capacity.Moreover,5G also brings the millimeter wave(mm Wave)band into the application scope to further improve the speed and capacity.However,millimeter wave has not been widely used because of its enormous path propagation loss.Massive MIMO provides a feasible scheme for the successful application of millimeter wave,which can resist the propagation loss of millimeter wave by the high gain brought by deploying a large number of antennas.In the massive MIMO technology,the channel estimation and precoding are the key technologies that affect the system performance.However,the deployment of a large number of antennas increases the complexity of the system,and the use of millimeter waves shortens the channel coherence time,which introduces great challenges to the channel estimation and precoding.Existing channel estimation and precoding schemes are difficult to be applied to millimeter-wave large-scale MIMO technology because of the high computational consumption or the low accuracy.Therefore,these two key technologies are studied in this thesis.With the increase of the number of antennas,the channel matrix becomes larger,which makes it difficult for existing algorithms to meet the requirements of low computational consumption and high accuracy.Therefore,a new channel estimation algorithm is proposed in this thesis.The algorithm formulates the problem of channel estimation as a problem of the recovery of sparse channel matrix,and utilizes the low rank property of the channel matrix in antenna domain as the side information to participate in the matrix recovery,which is modeled as a convex optimization problem.Then,the problem is solved by using the Augmented Lagrangian based Alternating Direction Inexact Newton(ALADIN).In addition,the characteristics of millimeter wave massive MIMO beams and the favorable property of propagation are taken as prior information in this thesis,which are applied to the proposed ALADIN based channel estimation algorithm.And then a channel estimation algorithm based on amplitude selection is proposed.Simulation results show that the two algorithms are of lower computational consumption and can improve the estimation accuracy,among which the latter algorithm has better performance at higher SNR and longer training length.Precoding technology and channel estimation are faced with similar difficulties.Traditional MIMO antennas are few in number,and each antenna can be configured with a separate RF chain to connect with the baseband digital precoder,so as to obtain the best system performance.However,with the increase of the number of antennas,the high power consumption and hardware complexity brought by the increase of the number of RF chains,make the high-performance digital precoding scheme unsuitable for the application in practice.At present,the mainstream scheme is the analog/digital hybrid precoding.This kind of scheme is realized with low power consumption and low hardware complexity in exchange for less loss of spectrum efficiency,but it also introduces the problem of high computational consumption,that makes the existing algorithms difficult to meet the requirements.In this thesis,a new hybrid precoding algorithm is proposed.In this algorithm,the hybrid precoding algorithm design is modeled as a spectral efficiency maximization problem,which is finally transformed into a nonconvex optimization problem to solve the Euclidean distance between the hybrid precoder and the unconstrained optimal precoder.And ALADIN algorithm is used to solve this problem.Simulation results show that the algorithm can achieve a better approximation of the spectral efficiency of the unconstrained optimal precoder,and is of a lower computational consumption.