Research On Distributed Compressed Sensing For Underwater Acoustic Sparse Channel Estimation

Compared with wireless communication channels,underwater acoustic channel is highly complicated due to its hostile time-frequency-space varing characteristics.Communication under such complicated channel pose great challenges to the research community.Channel estimation can provide the channel characteristics for channel equalizers,so that the communication performance will be improved.As underwater acoustic channel exhibits typically sparse,channel estimation can be transformed to a problem of sparse reconstruction under the framework of compressed sensing methods(CS).However,the performance of the convention channel estimation algorithms such as least square(LS)and orthogonal matching pursuit(OMP)will seriously degrade under short observation length and low signal-to-noise ratio(SNR).In this paper,in order to improve the performance of underwater acoustic channel estimation under short observation length and low SNR,and based on the coherence between multiple channels,we apply the distributed compressed sensing(DCS)to underwater acoustic channel estimation to simultaneously estimate channels.Thus,the performance of channel estimation will be improved under short observation length and low SNR,also the communication robustness and the bandwidth efficiency will be improved..In order to improve the underwater acoustic communication performance,we focus on channel estimation and channel equalization,especially distributed compressed sensing underwater acoustic channel estimation.The contribution of the paper is below:1.Joint spatial sparsity.In the SIMO and MIMO communication systems,multipath structures exhibit correlation.In this paper,joint spatial sparse distributed compressed sensing method was proposed for underwater acoustic channel estimation.For SIMO channel estimation,joint sparse model-2(JSM2)was utilized,the model for joint spatial sparse distributed compressed sensing was established,and the simultaneous orthogonal matching pursuit(SOMP)was utilized to optimize the model.For MIMO channel estimation,we established a new channel model which reconstructed the measurement matrices,and we proposed differential orthogonal matching pursuit(DOMP)algorithm to optimize the channel model,so that the co-channel interference(Col)was suppressed.Simulation results and filed experimental results showed the effectiveness of the proposed methods.2.Joint temporal sparsity.The multipath structures in adjacent data blocks show strong correlation.In this paper,joint temporal sparse distributed compressed sensing method was proposed for underwater acoustic channel estimation.The proposed method was used to estimate the long-time delay underwater acoustic channels under short observation length.We established the joint temporal sparse distributed compressed sensing model under the framework of JSM2.We also used the SOMP algorithm to optimize the model.Both simulation and sea trial data demonstrated the effectiveness of the proposed method.3.Joint frequency domain sparsity.The multipath structures among all sub-bands exhibit strong correlation.In order to adopt such correlation to improve the channel estimation performance,we proposed joint band sparsity distributed compressed sensing method for underwater acoustic multiband channel estimation.In the paper,we applied multiple selection strategy to distributed compressed sensing.By the proposed multiple selection strategy and the distributed compressed sensing,the common delays were improved and the differential delays could be reconstructed correctly.Furthermore,we jointed both band sparsity and temporal sparsity to further improve the multiband channel estimation.The simulation and filed experimental results showed the effectiveness of the proposed method.