Wireless Transmission Technology Of Millimeter Wave Large Scale MIMO Systems

The proliferation of multimedia infotainment applications and high-end devices drives an explosive de-mand for high-data rate services.This brings significant challenges to future wireless communication systems.The next 5th generation(5G)mobile communication places more demands on capacity,power consumption and bandwidth.Therefore,the exploitation of millimeter wave(mmwave)bands ranging from 30GHz to 300GHz has been considered as a promising solution due to its abundant spectrum resource.However,the mmwave signals experience more severe path loss than the signals in conventional sub-6GHz frequency band-s.To combat it,multiple antennas are usually adopted in the mmwave communication system to achieve high beamforming gain.In this thesis,we research on some key technologies for mmwave massive MIMO com-munication systems.The main work and innovations of this thesis are as follows:Firstly,the development history of the wireless local area network(WLAN)and the research status of mmwave communication are introduced.Meanwhile,the transmission characteristics of mmwave communi-cation and its potential challenges in the present and future are discussed in detail.Secondly,several key technologies in MIMO systems,such as space division multiplexing,antenna diversity,channel estimation,signal detection and classical precoding,are introduced in detail.Meanwhile,this paper compares the channel estimation performance of the LS and the MMSE algorithm,and the spectrum efficiency of the system under the ZF,MRT and MMSE precoding methods.Moreover,this thesis introduces the significant advantages of large-scale MIMO compared with traditional MIMO in terms of physical charac-teristics and system performance.In view of the fact that there is severe pilot pollution in large-scale MIMO systems,three mainstream suppression or elimination methods are introduced.In addition,the practical appli-cation of large-scale MIMO technology in millimeter wave systems is analyzed,which provides the relevant theoretical basis for the follow-up research of key technologies in millimeter wave large-scale MIMO systems.Thirdly,we focus on the problems of training codebook design and channel estimation for mmwave communication systems with uniform planner arrays.This paper presents a new codebook optimization design method from the perspective of shaping the beam pattern to achieve uniform beam alignment performance with low overhead.Specifically,the codebook design is formulated as an optimization problem,where the overshoot of the beam pattern is properly constrained such that each training beam is closed to the ideal one with a flat magnitude response and a narrow transition band.Furthermore,on the basis of the optimized training codebook,this paper proposes a new formulation for the mmwave channel estimation problem.This formulation captures the sparse nature of the channel,and enables leveraging tools developed in the adaptive compressed sensing(CS)field to design efficient estimation algorithm for mmwave channel.Simulation results show that the proposed training codebook can achieve a high accuracy of channel estimation with low error rate even in the low SNR condition.Finally,a geometric construction based hybrid precoding algorithm with low-complexity is investigated for mmwave communication systems.After transforming the original problem into an algebraic problem of matrix decomposition,we observe it in the complex plane from the perspective of geometric construction and provid a sufficient condition for the construction to be feasible.Then a smart designed baseband precoder is given to satisfy this condition and a triangle construction based matrix decomposition algorithm is proposed to complete the hybrid precoding design.To make the proposed algorithm optimal,the number of RF chains needs to be no less than twice the user number.However,a modified hybrid structure can relax this constraint by adding some low cost phase shifters without any performance loss.Moreover,the computation complexity of the proposed algorithm is much lower than that of the popular used OMP algorithm.Simulations validates the superior performance of the proposed algorithm.