Research On Channel Estimation Of Underwater Acoustic OFDM System Based On Compressed Sensing

Orthogonal frequency division multiplexing(OFDM)is a multi-carrier transmission technology,which has been widely used in underwater acoustic(UWA)communication systems to improve transmission speed and spectrum efficiency.The channel estimation technique improves the accuracy of data demodulation at the receiver by estimating channel state information(CSI).However,a large number of sub-carriers are required to transmit pilot information to ensure the accuracy of channel estimation due to the complexity of the UWA channel,which results in significant underutilization of spectrum resources.Due to the sparsity of UWA channels,the spectrum efficiency and estimation performance can be improved by using compressed sensing(CS)in channel estimation.However,most of the commonly used reconstruction algorithms in channel estimation requires channel sparsity as a priori information,which seriously restricts the application of these algorithms.Especially,due to the serious delay spread and Doppler spread of UWA channel,there are many columns of measurement matrix in channel estimation,which leads to high complexity.Therefore,this paper studies the channel estimation based on CS as follows:1.It has been found that the time-varying characteristics of the slow time-varying UWA channel with good channel conditions are not obvious.Therefore,a good estimation effect can be obtained by ignoring the influence of inter-carrier interference(ICI)in an OFDM symbol period.Based on the above reasons,this paper proposes a sparsity adaptive CoSaMP based on a dynamic threshold and weak selection of atoms(DW-SACoSaMP)based on the BELLHOP UWA channel model and the ICI-ignored OFDM channel estimation model.The DW-SACoSaMP algorithm adopts variable step size strategy to solve the problem of sparsity dependence in CoSaMP algorithm,avoidsthe problem of over-estimation caused by fixed step size,and improves the estimation accuracy.Moreover,the DW-SACoSaMP algorithm uses the dynamic threshold to reduce the number of the inner product calculations,adopts the atomic weak selection strategy based on the constraint of the number of atoms to optimizes the support set on the premise of ensuring the estimation accuracy.Simulation results show that compared to other reconstruction algorithms,the DW-SACoSaMP algorithm achieves higher accuracy of UWA channel estimation with less reconstruction time in the case of unknown channel sparsity.2.The fast time-varying UWA channel with bad channel conditions is so time-varying that the influence of inter-carrier interference can not be ignored,which makes the channel matrix a full matrix and increases the difficulty of channel estimation.Therefore,this paper proposes a fast OMP algorithm based on optimized iterative process(OIP-FOMP)based on parameterized time-varying underwater acoustic channel model and an ICI-aware OFDM channel estimation model.In this algorithm,the method of calculating the candidate path signature Hermitian inner product matrix in advance is used to avoid a lot of inner product calculations,the effective QR decomposition is used to solve the path amplitude instead of the complex LS method,and the problem of reconstruction failure caused by inaccurate delay selection is avoided by optimizing the Hermitian inner product matrix.Theoretical analysis and simulation results show that,compared with the OMP algorithm,the complexity of the OIP-FOMP algorithm is reduced by about 1/4 on the premise of ensuring the estimation accuracy.