Research On Iterative Reweighted Channel Estimation And Hybrid Precoding Algorithms For Millimeter Wave MIMO Systems

With the development of information technology and the development of Internet + transitivity,wireless communication equipment is increasing exponentially.The demand for highspeed and large capacity transmission is more urgent in the future.The development of millimeter wave Massive MIMO communication technology will bring fundamental changes to solve this problem.Increasing transmission rate,increasing antenna array gain and reducing path attenuation will be effectively improved in the future communication field.However,in millimeter-wave Massive MIMO system,a large number of antennas and radio frequency(RF)links at the base station increase the complexity of the system and greatly increase the energy consumption of the system.At the same time,large-scale base station antennas also increase the complexity of channel estimation.Aiming at the problems of channel estimation and hybrid precoding,this paper studies channel estimation and hybrid precoding.The main work of this article is as follows:(1)For the traditional channel estimation based on compressed sensing,the resolution loss is serious and the estimation of channel state information is not accurate,this paper proposes a superresolution channel estimation based on iterative reweight(IR).By optimizing the objective function through the Padam,the proposed scheme can iteratively make the AoAs/AoDs approach towards the optimal solution,and finally realize the super-resolution channel estimation.In the optimization,the sparsity and data fitting errors are controlled by the weight parameter.Simulation results show that compared with OMP channel estimation scheme and adaptive code channel estimation scheme,the proposed scheme has better performance such as normalized mean square error and spectral efficiency.(2)In the channel estimation scheme,Padam is introduced into the channel estimation,and the optimal solution of AoAs/AoDs is obtained when the objective function is minimized by Padam optimization.(3)For the hybrid precoding in millimeter-wave Massive MIMO systems,the codebook size is too large and the feedback cost is too high,then a codebook design based on angle of departure(AoD)is proposed,and a hybrid precoding algorithm based on angle-based uniform quantization(UQ-A)is derived from the codebook.The algorithm mainly uses the AoD constant in the angle coherent time to get codebook,then the codebook quantizes the channel vector,calculates the analog and digital precoding matrices,and calculates the receiver synthesizer matrix by the singul-ar value decomposition of the equivalent channel matrix generated by the two matrices.The simulation results show that the proposed algorithm under the premise of substantially reducing the RF chain can achieve spectrum efficiency very close to the performance of full digital precoding algorithm.In a single cell millimeter wave Massive MIMO system,Padam based on in-depth learning is used to obtain the optimal angle of arrival/departure by iteratively re-weighting,so as to obtain accurate channel state information.Hybrid precoding based on angular uniform quantization not only reduces codebook and feedback overhead,but also achieves spectrum efficiency performance close to full digital precoding.