Channel Estimation For Massive MIMO-OFDM Systems Based On Matrix Recovery

Orthogonal Frequency Division Multiplexing(OFDM)technology is one of the key technologies in 4G and 5G.In OFDM systems,coherent demodulation of received signals requires the support of Channel State Information(CSI).In Massive MIMO-OFDM systems,not only does the receiver need to know CSI,but the sender also needs in order to achieve resource allocation,beamforming,and precoding technologies.Using compressed sensing(CS)technology for channel estimation can use a small number of pilots to obtain the time domain CSI,but the frequency domain CSI of the channel is needed.Therefore,this thesis proposes a channel estimation algorithm based on matrix recovery(MR)theory for OFDM systems and Massive MIMO-OFDM systems.The research content of this thesis is as follows:Firstly,this thesis investigates the channel estimation problem of OFDM systems based on MR theory.The frequency domain channels of consecutive multiple OFDM symbols are represented in matrix form,and the Fourier transform matrix can be regarded as the generating basis of the matrix.The channel vector of a single OFDM symbol is sparse,so the channel matrix is low rank.By utilizing this property,the problem of minimizing the rank of the channel matrix is transformed into a problem of minimizing the kernel norm.The Singular Value Thresholding(SVT)algorithm in MR theory is used to restore the channel matrix.Simulation results show that compared with traditional channel estimation algorithms,the proposed algorithm can obtain highprecision channel estimation with fewer pilots.Compared with CS algorithm,the estimation accuracy is similar,but the frequency CSI can be obtained directly.Secondly,this thesis investigates the channel estimation problem of Massive MIMOOFDM systems based on MR theory.It considers the joint recovery of CSI on the base station side.After receiving the pilot signal sent by the base station,users estimate it separately,and then feedback the results to the base station side for joint channel estimation.Firstly,the frequency domain channel between multiple transmitting antennas and multiple single antenna users is represented as a low rank matrix form.So,the rank minimization problem of the channel matrix can be transformed into a kernel norm minimization problem of the channel matrix,and then solved using the SVT algorithm in MR theory.The simulation results show that the proposed algorithm can achieve high-precision frequency domain channel estimation in Massive MIMO-OFDM systems with fewer pilots.