Font Size: a A A

Research On Hybrid Beamforming In Massive MIMO Systems

Posted on:2020-07-21Degree:DoctorType:Dissertation
Country:ChinaCandidate:C HuFull Text:PDF
GTID:1368330590458910Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
Massive MIMO is the key technology in 5G.However,the energy consumption of RF chains at millimeter wave frequencies is still-exorbitant,and the conventional fully digital beamforming schemes appears to be unfeasible in massive MIMO systems.Since the full RF configuration requires one RF chain per antenna element,the fully digital beamforming schemes will result in a high cost in massive MIMO systems and overburden power consumption of RF chains at mmWave frequencies,which is contrary to the "Green Communication" in 5G.In order to make 5G a reality,the hybrid beamforming technology is proposed as the requirement of reducing number of RF chains.In this dissertation,we take the research on hybrid beamforming technology and propose two low-cost and high-performance hybrid beamforming schemes,aiming at the spectral efficiency of fully digital beamforming with reduced RF chains.In addition,the second scheme is extended to partially-connected subarray structure and overlapping subarray structure.The main work and innovations of this dissertation include the following aspects:1)The successive optimization of analog precoder and combiners in hybrid beamforming technology is studied,and a hybrid beamforming scheme with the analog precoder and combiners optimized successively is proposed.With the help of information theory,the design of hybrid beamforming is decoupled into two-stage optimization problem.In the first step,we aim to find an equivalent baseband channel by maximizing its capacity through the analog RF processing.In the second step,the inter-user interference is eliminated by the BD technique.Specifically,by expanding the expression of water-filling capacity of equivalent baseband channel,it is found that the product of eigenvalues,which are the gains of eigenchannels,is highly dependent on the analog processing.Thus,define the objective function of analog part by the product of eigenvalues.The analog combiner for each user is independently designed by fixing RF precoder,and then the RF precoder is optimized by the given RF combiners.The objective function is solved by two methods: convex optimization and phase aligning.2)The joint optimization of analog precoder and combiners in hybrid beamforming technology is studied,and a hybrid beamforming scheme with the analog part jointly optimized is proposed.Differing from the common idea that optimize a matrix by fixing the other one,this dissertation goes a further step to propose a hybrid beamforming scheme with the joint optimization of RF precoder and RF combiners.Firstly,it is concluded that the side with more RF chains will be inefficient.Thus,we assume that the number of RF chains at each side is equal,and the objective function turns into the determinant of the equivalent channel.To solve this new problem,we decompose the two-matrices joint optimization problem into multiple two-vectors joint optimization subproblems.The results show that if the number of RF chains is equal to the number of streams,the algorithm can achieve its maximum spectral efficiency.Therefore,such a low-cost and high-performance hybrid beamforming scheme will be useful in massive MIMO systems.3)The hybrid beamforming technology with partially-connected structures and overlapping subarray structures is studied.Besides,a hybrid beamforming scheme with non-fully-connected structures is proposed.In order to realize 5G "green communication",this dissertation go a further step to save cost and power by non-fully-connected structures,and extends the analog joint optimization scheme to partially-connected structures and overlapped subarray structures.Specifically,apply the constraints resulted by new structures to the objective function,and obtain the analog processing matrices by isolating the contribution of each pair of RF chains to the objective function.The results show that the non-fully-connected hybrid beamforming scheme has considerable advantages in spectral efficiency and energy efficiency,compared with the existing algorithms;the scheme can not only guarantee high spectrum efficiency but also high energy efficiency,which is very promising in 5G "Green communication".
Keywords/Search Tags:massive MIMO, hybrid beamforming, decoupled iteration, joint iteration, partially-connected structures, overlapped subarray structures
PDF Full Text Request
Related items