Research On The Key Technology Of Millimeter Wave Communication System Based On Hybrid Beamforming

Millimeter wave(mmWave)communication,as a research hotspot in recent years,has received extensive attention from researchers and industry.Since millimeter wave has rich spectrum resources,millimeter wave communication technology has become one of the core key technologies for the fifth generation mobile communication(5G)to improve the spectrum efficiency of the system.The wavelength of millimeter wave is very short,and more antenna units can be deployed in a smaller antenna panel space.Therefore,millimeter wave communication also enables the development of large-scale multiple-input multiple-output(MIMO)technology.Hybrid Beamforming(HBF)in massive MIMO systems can provide very high antenna array gain,which can in turn compensate for the shortcomings of mm Wave high path loss.Therefore,mmWave communication based on hybrid beamforming is the top priority of future mobile communication development.In view of this,this article studies the problems faced by the hybrid beamforming technology in mmWave communication systems from different perspectives.This paper first studies the problem of dimension-deficient channel estimation due to the lack of the signal dimension of the HBF architecture.Second,it studies the design problem of the fast HBF algorithm in high-speed mobile environments.Finally,it studies the antenna mutual coupling effect modeling and baseband solution based on the HBF architecture decoupling algorithm design problem.The main results and contributions of this article can be summarized as follows:(1)First of all,this paper studies the dimension-deficient channel estimation algorithm for the signal under-dimensional problem brought by the HBF architecture.Based on the single-user MIMO system,this paper mainly considers the signal dimension-deficient of the HBF architecture,and uses the sparsity of the mm Wave channel to redesign the channel estimation algorithm.Because the compressed sensing technology can recover the sparse signal better,this paper first converts the mm Wave channel from the spatial domain to the angle domain to obtain a virtual channel coefficient matrix,which further highlights the sparseness of the mm Wave channel.Secondly,in order to construct a general form of compressed sensing algorithm,this chapter performs a series of processing on low-dimensional received signals.In order to ensure the maximum recovery of channel state information and meet the Restricted Isometry Property(RIP)required by the compressed sensing algorithm,this chapter uses the received signal to construct an adaptive sensing matrix,so that the sensing matrix itself can contain a certain degree of effective channel information As far as possible,the sparse channel of the transceiver antenna dimension is recovered from the baseband received signal of the radio frequency(RF)link dimension.Due to the virtual channel representation method,the problem of leakage of adjacent angle channel information is generated.Therefore,in this chapter,the compressed sampling matching tracking algorithm is used to restore the sparse channel,and the main angle channel information and adjacent angle channel information of the virtual channel are retained to the greatest extent.Finally,the original mm Wave channel is restored using the inverse transformation from the angle domain to the space domain.This paper evaluates the effect of different radio frequency links and different signal-to-noise ratio on the algorithm performance.Simulation results show that with the increase in the number of radio frequency links,the performance of the proposed channel estimation algorithm is better than the comparison algorithm,and the spectral efficiency is very close to the performance of all-digital beamforming.(2)Secondly,for the high-speed mobile environment,especially the high-speed railway wireless communication environment,this paper studies the scheme design of fast HBF.In the high-speed mobile environment,due to the rapid change of the relative positions of the user and the base station,the problem of misalignment of the beam pointing is extremely likely to occur.This problem will greatly increase the probability of interruption of the communication system,reduce the quality of service(QoS)of the communication system,and make the user experience worse.Therefore,this paper proposes a multi-user HBF algorithm design scheme based on fast beam searching.Firstly,combining the received signal power and the error function,the expression of the upper bound of the interruption probability is derived.Then,the fast HBF algorithm is designed in two steps.The first step uses a stepwise refinement beam search algorithm with low searching complexity to obtain simulated beamforming pairs.The second step uses the properties of the generalized Rayleigh quotient based on the zero forcing(ZF)algorithm and the spectral efficiency maximization criterion.Optimal digital beamforming matrix.Finally,this chapter also analyzes the effect of channel estimation error on the performance of the algorithm.Simulation results show that the proposed algorithm reduces the interruption probability to a great extent;compared with other HBF algorithms,the spectral efficiency of the proposed algorithm is closer to that of all-digital beamforming.In addition,the appropriate beam switching period can also reduce the interruption probability to a certain extent and improve system performance.(3)Finally,for the mutual coupling effect of antennas,this paper studies the mutual coupling effect modeling method based on deep learning and the baseband decoupling algorithm with HBF architecture.In the hyper-scale MIMO system,the contradiction between the increasing number of antennas and the requirement for antenna miniaturization will cause the antenna mutual coupling effect to become more and more serious.Unlike traditional mutual coupling modeling methods,this article uses deep learning methods to model antenna mutual coupling effects.The basic idea of modeling is to use a large number of antenna input and output signals as the training data set of the deep neural network,and to predict the mutual coupling parameters by learning the model parameters of the antenna mutual coupling effect.Simulation results show that on the one hand,the larger the training data module,the larger the mean square error(MSE)of the prediction parameters,and the smaller the module,the smaller the MSE;on the other hand,the larger,the faster the MSE curve drops.However,the learning rate of the neural network cannot be improved without limitation,so it is necessary to jointly optimize the learning rate and the size of the training module in order to achieve the best learning effect.Based on the learned antenna mutual coupling matrix,this chapter studies the baseband decoupling algorithm with HBF architecture.Simulation results show that the proposed decoupling algorithm has a low bit error rate,which can alleviate the effect of antenna mutual coupling effect on the system bit error rate.