Research On Key Technologies Of Massive MIMO Systems Based On Hybrid Beamforming

Multiple-input multiple-output(MIMO)and millimeter wave(mmWave)communication are the two key technologies for the 5th Generation Mobile Network(5G),and the combination of the two can effectively increase the capacity of the system.Hybrid beamforming(HBF)is the core technology to promote the practical application of both in the 5G system.In view of this,this paper studies the HBF technology for massive MIMO systems,not only develops a more energy-efficient HBF architecture but also designs an effective HBF algorithm for different massive MIMO scenarios.Since the design of the HBF algorithm requires the channel state information(CSI),this paper also studies the channel estimation technique of the mm Wave MIMO system based on HBF.The main contributions of this paper are as follows:(1)Considering the four factors of system spectral efficiency,energy efficiency,the complexity of implementation and hardware consumption,this paper proposes a new hybridly-connected HBF structure.Meanwhile,based on the structure,this paper designs the HBF algorithm for a single-user massive MIMO system.Since the fully-and partially-connected structures are the two special forms of the hybridly connection HBF structure,the proposed HBF algorithm is equally applicable to the above structures.The simulation and theoretical analysis results show that when the number of radio frequency links is the same,the spectral efficiency,hardware consumption and implementation complexity of the single-user massive MIMO system based on the hybridly-connected architecture are between the partially-and fully-connected architectures,and the energy efficiency is better than the above two architectures.Moreover,the spectral efficiency of the proposed HBF algorithm is very close to that of the system where the analog precoding matrix is unconstrained.(2)For multi-user massive MIMO systems,this paper proposes a new HBF algorithm based on signal-to-leakage-and-noise ratio(SLNR)maximization.The algorithm considers decoupling the digital and analog precoder matrices,and then designing the digital and analog precoding matrices in two stages,respectively.In the first stage,the analog precoding and combining matrices are designed according to the maximization of the radio frequency link gain between the base station(BS)and users,and the closed-form solution of the analog precoding or combining matrix is obtained by minimizing the Euclidean distance between the unconstrained and constrained analog precoding or combining matrices.In the second stage,when the analog precoding and combining matrices are fixed,the digital precoding matrix is designed according to the SLNR maximization,and the closed solution of the digital precoding matrix is obtained by using the properties of the generalized Rayleigh quotient.Furthermore,based on the proposed HBF algorithm,the paper also analyzes the asymptotic performance of multi-user massive MIMO systems and the robustness of the algorithm to channel estimation errors.The theoretical and simulation results show that the performance of the proposed algorithm based on SLNR maximization outperforms the commonly used HBF algorithm,and it is robust to channel estimation errors.(3)The main problem in the design of HBF algorithm for multi-user massive MIMO orthogonal frequency division multiplexing(MIMO-OFDM)system is how to design suitable common analog precoding and combining matrices to maximize the spectral efficiency of the system.In view of the above problem,when the channel state information(CSI)is known,this paper develops two effective algorithms to design the common analog precoding and combining matrices based on channel average and tensor unfolding.When the analog precoding and combining matrices are fixed,since the digital precoding and combining matrices corresponding to all subcarriers are independent of each other,this paper designs the digital precoding and combining matrices of each subcarrier according to the SLNR maximization and minimum mean square error(MMSE),respectively.Considering that there is a certain error in the channel estimation for the actual system,this paper analyzes the robustness of the proposed algorithms to the channel estimation error.The simulation results show that compared with the existing algorithms,the proposed algorithm can obtain higher spectral efficiency,and the tensor unfolding algorithm is better than the channel average algorithm,but the channel average algorithm is more robust to channel estimation error.(4)In this paper,based on the sparsity of the mm Wave channel,the channel estimation problems for single-user and multi-user mm Wave MIMO systems are transformed into subspace fitting problem.Furthermore,considering the accuracy of channel estimation and computational complexity,this paper proposes three different schemes to solve the above subspace fitting problem.The first scheme requires a two-dimensional search in the candidate angle space,which is very complicated,but highly accurate.The second scheme converts the two-dimensional search of the angle space into a one-dimensional search and then estimates the direction of the signal transmission path by the orthogonal matching pursuit(OMP)algorithm.The accuracy of the scheme is affected by the path correlation,but the computational complexity is low.The third scheme combines the advantages of the above two schemes,and the accuracy and complexity are between the above two schemes.Furthermore,in order to improve the signal-to-noise ratio(SNR)of pilot signals,this paper develops the pilot beam pattern design scheme for massive MIMO systems adopting different HBF architectures.Theoretical analysis and simulation results show that the proposed channel estimation algorithms based on subspace fitting can obtain better channel estimation accuracy while reducing computational complexity.