Precoding Techniques And Performance Evaluation For MU-MIMO Systems In 5G Mm Wave Communications

To meet with the requirement of system capacity and data transmission rate in internet-of-things(IoT),a multi-user multiple-input-multiple-output(MU-MIMO)at millimeter wave(mmWave)band is getting great attention in coming fifth generation(5G)radio systems.Therefore,finding new algorithms with low cost that can be implemented to accommodate different types of applications is a hot issue.Accordingly,this thesis focuses on how to compensate for the huge mmWave path loss by using different techniques,in such a way we can increase the number of users with the acceptable service level and quality of service.Furthermore,the techniques in this thesis are based on designing a precoding/combining codebooks for MU-MIMO and massive MU-MIMO antenna systems as well as finding solutions for the challenges and issues faced in the 5G system and beyond.The main work and contributions in this thesis include the following five parts.Firstly,in this thesis,the background of MIMO technology is summarized and discussed in detail.In this regard,the motivations and advantages of beamforming including the key benefits are investigated.Besides,it also highlights the free space path loss(FSPL)which is considered a substantial path loss in mmWave band signal for MU-MIMO.In addition,the MIMO precoding and combining techniques are regarded as the core work in the thesis.Secondly,in this thesis,based on a fixed weight quantized beamforming codebooks,the conventional codebooks are evaluated regarding the array factor(AF)and signal to noise ratio(SNR)as a common metrics.Besides,a new structure of quantized codebook is constructed and compared with the existing codebooks.In addition,as known that a mmWave radio system can operate efficiently when the massive number of antenna elements deployed in MIMO,especially in the base station(BS).Therefore,the effect of a number of antenna elements on the beamforming is addressed,and the ability to cover multiple users by generating directed narrow beams with a significant gain and directivity is investigated as well.Thirdly,by characterizing a three-dimensional(3-D)massive MU-MIMO antenna in mmWave communications,and by adopting matrix process based on CUR-decomposition technique and beam steering codebook,a novel algorithm is designed to solve the dimensionality problem caused by deploying massive antenna elements at the BS end.Explicitly,simulation results show that the CUR-decomposition algorithm is expected to achieve a significant system performance with low cost.Furthermore,the correlation between the number of antenna elements and a number of feedback bits(channel state information feedback bits)required by the massive MU-MIMO system is investigated and validated.Fourthly,based on a signal-to-leakage-plus-noise ratio(SLNR)instead of the commonly used signal-to-interference-plus-noise ratio(SINR),the statistical eigen-mode(SE)and zero forcing(ZF)beamformer models in Ricean fading channel are examined to achieve linear achievable sum-rate capacities.Besides,the dirty paper coding(DPC)method is discussed to achieve non-linear capacity form regarding the number of antenna elements and number of users.Moreover,in the thesis,a novel algorithm to design a code weight vector for massive MU-MIMO system is developed,which is based on a power approximation method and SLNR to find the optimum solution of the weight vectors for all users simultaneously to get an available channel matrix which minimizes the co-channel interference(CCI)for all user ends.Furthermore,the average system capacity of the technique is analyzed and compared with the conventional method is known as singular value decomposition(SVD)method.Fifthly,the beam training cost and energy efficiency are studied.In this regard,a new beam training technique for MU-MIMO mmWave communications system is studied,which is based on a single radio frequency(RF chain)sequential beam steering codebooks to serve multi-users sequentially by deploying only full analog precoder/combiner,in which a single RF-chain is deployed at the BS and MS.Moreover,to investigate the system effectiveness when implementing a single RF chain at the base station(BS).new arbitrary beam training functions are proposed regarding the increased number of users,the effectiveness is validated in terms of average system capacity.The proposed techniques and algorithms in this thesis,in particular for 5G mmWave massive MIMO system are expected to be validated and implemented in coming 5G radio systems to enhance the system performance.