Research On Channel Estimation Method For Underwater Acoustic Communication Based On Filter Bank Multicarrier

Compared with the terrestrial wireless channel,the underwater acoustic channel is faced with more severe limitations,mainly manifested in the complex time-varying,frequency-varying,space-varying characteristics,serious multipath interference,scarce spectrum resources and large signal transmission loss.FBMC uses a filter bank which is composed of a prototype Filter with good time-frequency focusing performance to perform pulse shaping filtering on the up-sampled signal.Compared with OFDM,Frequency spectrum utilization was improved on the basis of the capability of enhancing Inter Carrier Interference?ICI?and Inter Symbol Interference?ISI?.On this basis,FBMC technology adopts Offset Quadrature Amplitude Modulation technology to further resist ICI and ISI without increasing the cost of spectrum resources.These advantages make FBMC technology more suitable for underwater acoustic environment than OFDM technology.However,FBMC technology also has some problems.The subcarrier basis function only maintains the orthogonal property in real domain,which will introduce the inherent imaginary part interference between symbols and subcarriers,thus causing some difficulties in channel estimation and equalization of communication system.Therefore,it is very important to study the high precision channel estimation algorithm for FBMC to improve the performance of underwater acoustic communication system.To solve this problem,this paper mainly proposes two channel estimation methods based on FBMC.Firstly,a FBMC channel estimation method based on the improved Recursive Least Squares?RLS?algorithm is proposed to solve the problem of high complexity and low accuracy when using the traditional adaptive algorithms to estimate the sparse channels.The first step is inserting the training sequence into the frame structure of the FBMC transmission signal.Then,process the data before and after the underwater acoustic channel and estimate the channel impulse response gradually through adaptive filter method.The proposed algorithm uses the sparsity of underwater acoustic channel to add penalty items to the cost function of RLS algorithm to constrain the channel coefficient.On the one hand,the cost function is processed by sliding window.On the other hand,method of binary coordinate descent is applied to searching the optimal solution.Through the simulation analysis,the channel estimation accuracy of the proposed algorithm has reached 10^-3 orders of magnitude and its complexity is lower than the traditional algorithm.The simulation results validate the the superiority of the proposed algorithm in the accuracy and complexity of underwater acoustic channel estimation compared with the traditional algorithm.Since the sparse adaptive channel estimation method based on training sequences brings additional spectrum resource overhead.This paper proposes the FBMC discrete pilot channel estimation method based on compressed sensing according to the superior performance of compressed sensing technology in the reconstruction of sparse signals.In the pilot structure design,this paper designs the E-IAM-C discrete pilot structure to solve the problem of inherent imaginary part interference in FBMC channel estimation based on the idea of Interference utilization and equivalent pilot energy maximization.In the channel reconstruction algorithm,this paper uses g OMP algorithm which is of high performance to reconstruct the underwater acoustic channel.The simulation analysis shows that the estimation performance of the proposed E-IAM-C discrete structure is better than the other discrete pilot structure and the proposed method only uses 63 pilots to reach better performance than traditional block pilot structure which saves nearly 63%of the spectrum costs.The proposed method not only ensures the underwater acoustic channel estimation precision,but also effectively saves the FBMC system spectrum.