Research On Sparse Channel Estimation Algorithm For OFDM System Based On Compressed Sensing

Orthogonal frequency division multiplexing(OFDM)system is widely used because it can effectively resist multipath effect.In particular,the ability of receiving equipment to accurately obtain channel state information(CSI)of OFDM system becomes very important.This thesis firstly studies the land wireless communication system,and improves the two reconstruction algorithms by means of the sparsity level of multipath channel;next the underwater acoustic(UWA)communication system is investigated.By focusing on the advantages of OFDM system,the sparsity level and complexity of UWA channel,the compressed sensing(CS)theory is used to reconstruct UWA channel,and an innovative reconstruction algorithm is selected.The main contribution of this thesis are as follows:Firstly,the sparse channel model is constructed according to OFDM and MIMO-OFDM systems,and the sparse representation of channel over delay domain is discussed.From the perspective of CS theory,the pilot which can reconstruct CSI with high probability is designed.Based on CS theory,the CSI is reconstructed by orthogonal matching pursuit(OMP)algorithm,OMP algorithm relies on the information of sparsity level.In this thesis,an automaticallystopped orthogonal matching pursuit(As OMP)algorithm is proposed,which can stop iteration automatically and performance better than OMP algorithm in channel estimation.By statistical analysis,the number of iterations of As OMP algorithm is almost the same as that of OMP algorithm.Secondly,the CSI is reconstructed by compressed sampling matching pursuit(CoSaMP)algorithm,CoSaMP algorithm makes the residuals converge by eliminating the wrong entries,occasionally,the convergence of the residual is uncertain,which leads to the unstable convergence rate of CoSaMP algorithm.In this thesis,a fast compressed sampling matching pursuit(F-CoSaMP)algorithm is proposed,which can guarantee the absolute convergence of residuals,the experimental results show that the performance of F-CoSaMP algorithm is better.Thirdly,the sparse channel model of UWA-OFDM system is constructed by combining the advantages of OFDM system with the sparsity of UWA system,prominent multipath effect and significant Doppler shift,the sparse representation of channel in Doppler delay domain is studied.The CSI is reconstructed by using stage-determined matching pursuit(SdMP)algorithm,compared with OMP algorithm and CoSaMP algorithm.By comparison,OMP algorithm and SdMP algorithm have better estimation performance,and their performances are similar;according to statistics,the running time of OMP algorithm is shorter when the sparsity level is low,and extends with the increase of the sparsity level,the running time of SdMP algorithm is good in most cases,which is suitable for complex UWA communication system.