Sparse Channel Estimation For Reconfigurable Intelligent Surface Assisted Millimeter Wave Massive MIMO System

With the technology maturity and commercial deployment of the fifth generation mo-bile communication system(5G),the development and research of the sixth generation mobile communication system(6G)technology has been widely concerned.Reconfig-urable intelligent surface(RIS)is considered as one of the potential technologies of 6G network.Relevant experimental results show that RIS can effectively control the phase,amplitude and frequency difference between the reflected signal and the incident signal,so as to intelligently control the scattering,reflection and refraction characteristics of ra-dio waves to help the wireless communication system overcome the negative effects in the natural wireless propagation environment.Therefore,many existing studies suggest using RIS technology to assist millimeter wave(mm Wave)masssive multiple input and multiple output(Massive MIMO)system to further improve the spectrum efficiency,sys-tem capacity,transmission reliability and system coverage.In order to make full use of the advantages of RIS- assisted mm Wave massive MIMO system,the system needs to obtain accurate channel state information(CSI).Currently,many sparse signal recovery algo-rithms based on compressed sensing are applied to channel estimation in the RIS-assisted mm Wave Massive MIMO system.However,most of the existing studies assume that the system is in ideal hardware conditions and ideal environment,while the impact of some non-ideal conditions on the system performance and the problem of channel estimation still need to be studied.In view of this,this dissertation focuses on the following issues:Firstly,this dissertation analyzes the outage performance of RIS assisted mm Wave Massive MIMO system under non-ideal CSI.This dissertation provides an analysis frame-work based on joint active and passive beamforming technology,and focuses on the outage performance when the channel parameters related to RIS are non-ideal.Based on double-Nakagami distribution and cosine antenna pattern model,the expression of system outage performance is derived.Based on convex optimization theory and Hermite-Hadamard inequality,the upper and lower bounds of outage probability are derived to estimate the range of outage probability with low complexity.Using a similar framework,the outage probability expression of traditional relay system is derived to compare with the outage performance of RIS system.Monte Carlo simulation results show that the analytical ex-pression can accurately fit the simulation results.Compared with the traditional relay system,it is proved that the robustness of RIS system to non-ideal CSI is poor.Secondly,this dissertation studies the channel estimation problem of RIS-aided mmWave Massive MIMO system with low-resolution ADC.In order to deal with the challenge of channel estimation brought by cascaded channel in mmwave massive MIMO system,this dissertation deduces the equivalent channel model based on cascaded channel model,and uses equivalent channel estimation to replace the estimation of cascaded channel.Ac-cording to the sparse structure of traditional mmWave channel in virtual angle domain,it is deduced that the equivalent channel is row--column block sparse in virtual angle do-main.On this basis,this dissertation proposes an EM--NNL--GAMP algorithm based on expectation maximization(EM)algorithm,nearest neighbor learning(NNL)algorithm and generalized approximate message passing(GAMP)algorithm.The algorithm can adaptively reconstruct the sparse structure in the virtual angle domain of the equivalent channel to improve the channel estimation performance and compensate for the impact of low--precision ADC quantization on the system performance.Monte Carlo simulation results show that the proposed algorithm can improve the channel estimation performance without increasing the computational complexity.Thirdly,the problem of array diagnosis and channel information joint estimation in RIS-assisted mm Wave Massive MIMO system is studied.Based on the transmission model of system with the RIS suffing array blockage,the impact of array blockage on power / Signal--to--Noise Ratio gain is analyzed.In this dissertation,we parameterize the cascaded channel of the system,and propose a two-stage estimation scheme based on this parametric channel model to complete the joint estimation of array diagnosis and channel parameters.In the first stage,iterative reweighted algorithm is used to estimate part of the channel parameters.In the second stage,batch algorithm and two-timescale online opti-mization algorithms are proposed to complete the array diagnosis and the estimation of the remaining channel parameters.On this basis,a noise reduction algorithm is proposed,which speeds up the convergence speed of the two methods and improves the estimation performance.Simulation results verify the effectiveness and performance advantages of the proposed algorithms.Finally,the problem of array diagnosis and channel information joint estimation in RIS--assisted multiuser mm Wave Massive MIMO system is studied.Considering the shortcomings of the parameterized channel model in multi-user scenarios,this disserta-tion derives the equivalent channel model based on cascaded channels.According to the derivation process of equivalent channel model and the joint sparse structure of the vir-tual angle domain channels between users and RIS,the sparse structure of multi-user equivalent channel in virtual angle domain is derived.Based on the sparse structure,a pilot-assisted joint array diagnosis and channel estimation scheme is proposed.For small number of users and large number of users scenarios,this dissertation proposes a joint orthogonal matching pursuit array blocking calibration(JOMPABC)algorithm and two low complexity two-space-scale optimization algorithms to realize the joint estimation of array diagnosis and effective channel information.Monte Carlo simulation results show that the proposed algorithms can achieve better channel estimation performance.