Research On Channel Estimation And Hybrid Precoding Technology In Millimeter Wave Communication System

With the ever-increasing demand for wireless data traffic,today's communication systems are faced with shortages of spectrum resources,high transmission delays,and low throughput.Millimeter wave(mm Wave)can effectively solve the problem of shortage of spectrum resources due to its frequency bandwidth.At the same time,the combination of millimeter wave technology and large-scale multi-input multi-output(Massive MIMO)technology can achieve an order of magnitude increase in system throughput.Therefore,millimeter-wave communication has received widespread attention in the research community.The millimeter wave communication system deploys large antenna arrays at the transmitter and receiver.Due to the mixed signal power limitation and the high frequency of the millimeter wave,new MIMO signal processing technology is required.Therefore,it is of great significance to study the hybrid precoding design and channel estimation in Massive MIMO millimeter wave communication systems.In this paper,under the background of millimeter wave communication,the research on hybrid precoding design and channel estimation is conducted.The main research results are as follows:1.Study the design of hybrid precoding in millimeter-wave Massive MIMO communication systems.Aiming at the problem that the multi-user hybrid precoding algorithm in millimeter-wave Massive MIMO downlink communication system has a high computational complexity,a hybrid precoding scheme based on gradient projection is proposed.First,the problem of solving complete digital precoding is expressed as an expression containing matrix inversion,and then the inverse of the matrix is approximated by the Neumann series sum to reduce the complexity of obtaining complete digital precoding;On the basis,the alternating projection minimum algorithm of gradient projection is used for hybrid precoding design,and the complete digital precoding matrix is decomposed into analog precoding matrix and digital precoding matrix.The gradient projection algorithm is used in the simulation precoding design part,which reduces the computational complexity while ensuring the algorithm convergence.Simulation results show that,compared with the existing hybrid precoding design based on the alternating minimization algorithm,although the proposed algorithm has a slight loss in rate performance,it reduces the computational complexity.2.Study the channel estimation problem in millimeter-wave Massive MIMO communication systems.Aiming at the problem of low accuracy of channel state information estimation in millimeter wave channel environment,a channel estimation scheme based on deep learning is proposed,which transforms the channel estimation problem into a sparse signal recovery problem under multi-measurement vector(MMV).The sparse paths in the millimeter wave channel have correlations,and the sparse structures under different sparse paths are different,so consider using a recurrent neural network(RNN)with long and short-term memory(LSTM)units to effectively capture the correlation between sparse channel paths Sexual and sparse structure.The offline training of the channel data obtains the parameters of the cyclic neural network.The theoretical analysis and simulation results show that the proposed algorithm can improve the channel estimation accuracy compared with the traditional MMV-based channel estimation algorithm.