Research On Channel Estimation For Underwater Acoustic OFDM Communication System Based On Compression Sensing

The development of underwater acoustic communication technology is seriously affected by the complexity of multipath time delay,dual-choice characteristics of time and frequency and high noise.Within the limited bandwidth of underwater acoustic channels,Orthogonal Frequency Division Multiplexing(OFDM)can achieve higher bandwidth utilization and has the ability to resist multipath delay spread.Channel State Information(CSI)estimation in OFDM systems can be used to improve the accuracy of demodulation data.However,the traditional channel estimation methods have low accuracy in complex underwater acoustic environment,and need a large number of pilots to achieve good estimation results,which seriously occupies the spectrum resources.In view of the complexity and sparsity of underwater acoustic channel and the technical characteristics of OFDM communication system,this paper proposes an underwater acoustic channel estimation method combined with compression perception theory and pilot pattern optimization method to improve the accuracy of channel estimation,for better communication performance.The main work of the paper is as follows:1.Firstly,the underwater acoustic OFDM communication system is introduced,the principle and advantages of OFDM are studied,and the influence of some key technologies such as cyclic prefix,channel coding and channel estimation on OFDM communication system is analyzed.Then the characteristics of underwater acoustic channel environment are analyzed and the simulation of the intrinsic sound line and channel impulse response is completed.Finally,the performance of the system is compared with the OFDM system by simulation under different pilot insertion modes,the effectiveness of the system is verified.2.The channel estimation algorithm based on compression perception is studied.Firstly,this paper introduces the theory of compression perception and analyzes the advantages and disadvantages of some existing algorithms by simulation: matching pursuit algorithm has high reconstruction accuracy and low complexity,but it can not be used in the case of unknown sparsity;The adaptive matching pursuit algorithm can approach the true sparsity by updating residuals,but the initial step size of the algorithm will affect the reconstruction accuracy and quality,and the pseudo-inverse method of least squares is affected by noise.Therefore,this paper adopts an iterative scheme to change the step length stage by judging the current scale of the support set,which can control the expansion speed of the support set in different stages,and reduce the probability of over-estimation and under-estimation,the algorithm of coordinate descent is introduced to improve the anti-noise ability of sparse vector reconstruction.Through simulation analysis and pool experiment,the proposed algorithm performs better than the traditional algorithm when applied to underwater acoustic channel estimation.3.The optimization method of pilot pattern search is studied.In order to solve the problem of poor performance of channel estimation in the case of insufficient pilot number,several existing algorithms are compared and analyzed in this paper,then a tree-like random search algorithm based on the new criterion is proposed to design the pilot pattern,so OFDM communication system can still maintain good communication performance in the sparse underwater acoustic channel.The simulation and pool experiment results show that the proposed algorithm converges more easily than other algorithms when searching for the optimal pilot pattern in sparse channel,and the coherence coefficient of the corresponding criterion is smaller,when the corresponding pilot pattern is used for channel estimation,the mean square error is smaller and the system error rate is lower.