Research On Wireless Channel Estimation Based On Adaptive Compressed Sensing Reconstruction Algorithm

Compressed Sensing(CS)-based channel estimation is a research hotspot that has appeared in recent years.CS theory combines data collection and compression into one step,and completes sampling,transmission and storage of signals at a lower rate,which greatly saves transmission costs and storage resources,so it is widely used in signal processing.The Orthogonal Frequency Division Multiplex(OFDM)system has a high bandwidth efficiency,which is very suitable for scenarios where spectrum resources are scarce.Its channels are sparse due to the delay spread caused by multipath propagation.In large-scale Multiple Input Single Output(MIMO)systems,multiple antenna arrays can be used to obtain large spatial multiplexing gains,and improve system capacity and spectral efficiency.The channels in large-scale MIMO systems are also sparse in the angular domain due to scattering effects.The traditional channel estimation method obtains a certain channel estimation performance under the condition of using a large number of pilots.However,the traditional channel estimation method is no longer applicable when the channel bandwidth is expanded to a certain degree.In order to make better use of the sparse characteristics of the wireless channels to reduce the number of pilots,CS technology can be used to complete the channel estimation.A large number of documents show that CS-based channel estimation can obtain excellent estimation performance with only a few pilots.In this paper,the sparsity adaptive CS reconstruction algorithms are studied in depth and applied to sparse channel estimation in OFDM and massive MIMO systems,so that channel estimation can be completed with high quality without known channel sparsity.The main contributions of this article are as follows:1)An improved sparsity adaptive CS reconstruction algorithm—Weak Selection Stagewise Adaptive Matching Pursuit(WSSt AMP)algorithm is proposed.This algorithm first deletes the unsatisfactory atoms at the atomic primary selection stage by setting the fuzzy threshold.Deleting a large number of unsatisfactory atoms ensures that the selected atoms have a large correlation with the residuals,resulting in that the algorithm has a high reconstruction accuracy.In addition,for the case of variable step sizes at different stages,a power function type change is designed.The method of step size enables the algorithm to expand the support set faster to approximate the true sparsity in the large step size phase,which further reduces the reconstruction time of the algorithm.Simulation results show that the reconstruction performance of WSSt AMP algorithm is better than other algorithms regardless of the presence or absence of noise,and its algorithm complexity is lower.So it has more practical significance.2)The WSSt AMP algorithm proposed in this paper is applied to the sparse channel estimation in OFDM system,and its performance is compared with other algorithms.According to the simulation results,the proposed WSSt AMP algorithm still has very high channel estimation performance in actual channel estimation.As compared with reconstruction algorithms such as Sparsity Adaptive Matching Pursuit(SAMP),its reconstruction performance of mean square error(MSE)is improved by about 2-3d B;3)The WSSt AMP algorithm is applied to the channel estimation of massive MIMO systems,and the channel estimation performance is also compared with those of other algorithms.The change of the channel estimation performance of the WSSt AMP algorithm under different numbers of pilots and the number of transmitting antennas is also studied.Simulation results show that in large-scale MIMO systems,the MSE performance of WSSt AMP algorithm is still better than SAMP and other reconstruction algorithms,which is improved by 1.5-2.5d B.In addition,with the decrease of the the number of transmitting antennas or the increase of the number of pilots,the WSSt AMP algorithm can obtain more excellent reconstruction performance.