Research On Limited Feedback And Beam Tracking In Massive Mimo Systems

Massive multiple input multiple output(mMIMO)is one of the key technologies of the fifth generation wireless systems(5G).Deploying large-scale antenna array at the transmitter or receiver can improve system performance by increasing the spatial freedom of transmission channel.It can be divided into two kinds of mMIMO system according to the system frequency,namely the sub-6 GHz mMIMO system and the millimeter wave(MMW)mMIMO system.One important premise for mMIMO system to obtain high system gain is that the transmitter knows full or part of channel state information(CSI).However,the estimation or feedback of CSI of mMIMO system will lead to huge resource overhead.Therefore,the study of limited feedback of CSI or beam tracking technology of MMW mMIMO system is important to obtain the performance gain of mMIMO system.This thesis topic is supported by the research project "Research on Key Technologies of 5G Wireless Air Interface" cooperated with an enterprise.The problem of huge CSI feedback overhead in mMIMO systems is studied.For sub-6 GHz massive MIMO system,a spatio-temporal correlated channel feedback scheme which joint KLT(Karhunen-Loeve Transform)and polar-cap quantization is proposed.For MMW mMIMO systems,unscented Kalman filter(UKF)based beam tracking algorithm is proposed.The specific work is as follows:1)The technical challenges faced by mMIMO and various key technologied of mMIMO are reviewed.The mMIMO system can be divided into two different types according to the system frequency.For the systems work at sub-6 GHz,the research of limited CSI feedback is mainly discussed,while for MMW mMIMO systems,the study of beam tracking is mainly discussed.2)For sub-6 GHz maMIMO systems,a limited CSI feedback scheme which jointly performs KLT and polar-cap quantization is proposed for spatio-temporal correlated mMIMO channel.The scheme exploits the spatial correlation of the high-dimensional CSI and reduces the dimensionality of the high-dimensional CSI vector.Then a polar-cap quantization codebook is generated according to CSI time-correlation coefficient.As a result,this CSI feedback scheme can not only reduce the cost of CSI feedback but also improve the quantization accuracy of CSI.Simulation results show that the proposed scheme can achieve higher system sum rate performance than schemes that only consider the temporal correlation of channel or only consider the spatial correlation of channel to reduce the cost of CSI feedback.3)For the mobile MMW mMIMO systems,a beam tracking method based on UKF is proposed.The beamforming system of MMW mMIMO is nonlinear and the the nonlinearity is strong when the system is in mobility.Compared with traditional EKF algorithm,UKF algorithm can obtain higher estimation accuracy for nonlinear state estimation problems.Thus,the UKF algorithm is introduced into the beam tracking process of MMW mMIMO system for a moving scenario and it can effectively improve the beam tracking accuracy.Simulation results show that compared with the traditional beam tracking algorithm like EKF based on first-order Taylor expansion,the proposed beam tracking algorithm based on UKF can reduce the error of beam angle estimation to a lower level.