| With the development of 6G wireless system,Terahertz(THz)communication is regarded as an emerging technology in the research of next generation wireless systems.Due to its ability to provide larger spectrum resources and ultra-high transmission rate.The Reconfigurable Intelligent Surface(RIS)can control the phase shift of a single reflection unit and create a controllable communication environment.Therefore,the problem of insufficient coverage of multiple-input multiple-output(MIMO)communication is solved.The introduction of RIS has changed the traditional channel environment,and due to the passive characteristics of RIS,the traditional channel estimation method cannot be migrated well.Therefore,the acquisition of Channel State Information(CSI)is a key problem in RIS aided communication system.This thesis introduces the most common channel estimation schemes based on compressive sensing,and proposes a deep learning based channel estimation,an atomic norm minimization based channel estimation,and a channel estimation scheme based on adaptive grid matching tracking algorithm.For the channel estimation schemes raised above,the accuracy and computational complexity of each scheme are compared by simulation.Firstly,according to the sparse characteristics of THz channel,the channel estimation problem is transformed into a sparse recovery problem,and a channel estimation scheme based on compressed sensing sparse recovery algorithm is proposed.Furthermore,in order to reduce the high computational complexity brought by meshwork,a channel estimation scheme based on deep learning is further proposed.By building a three-segment neural network to fit the sparse reconstruction process,the channel estimation with low complexity is realized.Secondly,in order to solve the grid mismatch problem of the on-grid compressive sensing,this thesis studied the channel estimation of RIS-assisted THz communication system using atomic norm minimization algorithm based on the off-grid compressed sensing theory.It is include two stages,the first stage according to the received signal through atomic norm minimization algorithm to estimate base station AOD and the client AOA.In the second stage,the angular information estimated in the first stage is used to design the combination matrix and the pilot signal.Then Atomic norm minimization algorithm is used to estimate the channel gain and RIS Angle difference matrix.In the end,this thesis on the basis of based on meshless compression scheme,an adaptive grid matching pursuit algorithm,by integrating the concept of on-grid compressed sensing algorithm,according to the perspective of one phase of information,in the small scope set low lattice grid points,sparse signal recovery of low complexity,high resolution.The robustness of the algorithm is improved and the training cost of channel estimation is reduced. |
