Research Of Hybrid Pre-coding In Mmwave Massive MIMO Systems

The 3rd Generation Partnership Project(3GPP)defines three application scenarios for The 5th Generation Mobile Networks(5G).The enhanced mobile broadband(e MBB)scenario corresponds to the mobile broadband service with large traffic.Millimeter wave(mm Wave)and massive multiple-input multiple-output(Massive MIMO)technologies lay the foundation for ultra-high speed data transmission,which are significant support-ing technologies to realize e MBB.The low complexity downlink multi-user interference control is a very important research direction in the mm Wave Massive MIMO systems.Hybrid pre-coding technology,which can reduce the number of working radio frequency links,compromises the complexity and performance,and has gained widespread attention.At present,the research on hybrid pre-coding mainly considers the fixed array structure of analog domain and the limitation of constant amplitude,instead of fully consideration of the array structure and non-ideal characteristics of analog components,which affects the comprehensive evaluation and performance improvement of hybrid pre-coding.In order to solve the above problems,based on the antenna array system with the dynamic array and the Insertion Loss(IL)of analog components,considering the cen-tralized and distributed systems,this paper designs and analyzes the channel estimation scheme and hybrid pre-coding scheme,which are specifically divided into the following three aspects.Firstly,for the modeling and design problem of antenna array system,an antenna array system with the dynamic array and the IL of analog components is proposed.Based on the system transmission model,the performance balance is analyzed,and the demon-stration problem of system performance gain is solved.In addition,considering the rate service ratio,a distributed mm Wave Massive MIMO system is proposed,including system model and transmission protocol,in order to increase the achievable rates of users.Secondly,to solve the channel estimation problem,based on the antenna array sys-tem with dynamic array and IL of analog components,a transmission model of channel estimation is analyzed.Under the channel estimation transmission model,considering the system performance balance,a multi-cluster component information estimation scheme is proposed.It improves the average detection probability and solves the problem of high-precision channel estimation in centralized scenarios.Based on the location information,an estimation scheme of the beam state information(BSI)is proposed to avoid the increas-ing estimation complexity with the increasing number of distributed arrays.It solves the problem of efficient channel estimation in distributed scenarios.The simulation results verify the effectiveness of the two schemes.Finally,aiming at the design problem of hybrid pre-coding scheme,taking into ac-count the minimum rate requirement of each user and energy consumption constraint,the minimum mean square error(MMSE)hybrid pre-coding design problems in the central-ized and distributed scenarios are modeled,respectively.Considering the entire energy consumption constraint and the balance of system performance,two hybrid pre-coding schemes are proposed in centralized scenarios,namely the decoupling hybrid pre-coding scheme and the joint hybrid pre-coding scheme with higher complexity and higher perfor-mance.Both of them can improve the performance of bit error rate(BER)and sum-rate.The average BER,sum-rate,computation complexity and effect of estimation error of multi-cluster components on the signal to interference plus noise ratio(SINR)are ana-lyzed.Considering the each energy consumption constraint and the balance of system performance,a hybrid pre-coding scheme utilizing the difference of macro-diversity in the distributed scenarios is proposed to improve the system performance.The computa-tion complexity and the effect of estimation error of BSI are analyzed.Considering the spatial characteristics of the proposed distributed system,to solve the problems of high computation complexity and high requirement of BSI of the proposed hybrid pre-coding scheme in the distributed scenarios,a diagonalized simplified MMSE hybrid pre-coding scheme is proposed.The computation complexity and effect of estimation error of BSI are analyzed.The simulation results show that in the high signal to interference ratio(SIR)region,the performance of the diagonalized simplified scheme is close to that of the pro-posed former scheme utilizing the difference of macro-diversity,and it is not sensitive to the estimation error of BSI.