The Research Of The Channel Estimation Based On Sparse Sampling In Underwater Acoustic OFDM Communication System

Orthogonal Frequency Division Multiplexing (OFDM) can take full advantage of thelimited bandwidth of underwater acoustic communication (UWA) channel and resistinter-symbol interference effectively, which caused by multi-paths transmission. Meanwhile,it also can combine with Multiple-Input Multiple-Out-put (MIMO) to obtain the multifoldchannel capacity without additional frequency bands, which has a promising future in UWAcommunication. While with the increase of communication distance, the underwater channeldeteriorated and the influence of underwater acoustic signal caused by multi-paths andDoppler intensifies, which lead to the received signal distortion in phase, frequency andamplitude. So it needs to estimate the state of the underwater acoustic channel tocompensate the received signal. Since the advantages of simple calculation, fast rate ofconvergence and easily realization, the channel estimation based on pilot gets extensiveapplication. But the method is based on the rich multipath assumption, it need to insert alarge number of pilots to obtain accurate channel information, leading to low frequencyspectrum utilization. However, more and more physical arguments and experimentalevidence suggest that underwater acoustic channel appear sparse characteristics. It canlargely reduce the pilots and improve the spectrum efficiency by excavating the sparsecharacteristics of wireless channel deeply and taking it into full exploitation.The theory of compressive sensing (CS) or compressive sampling (CS) has been putforward in recent year, which break through the traditional signal acquisition method. Thetheory utilizes the sparse characteristics of signal, sampling and compressing at the sametime. Then we can reconstruct the original signal from relatively few samples efficiently. Sothe theory achieves great development in the areas of sparse signal reconstruction. Theinherent sparse characteristics of underwater acoustic channel makes the channel estimationcan be regard as the reconstruction problem of sparse signals, then using compressivesensing sampling and reconstructing the channel. So this paper in-depth analysis andresearch the underwater acoustic channel estimation based on sparse sampling.In this paper, based on the analysis characteristics of underwater acoustic channel, thebasic principle of OFDM and MIMO technologies, combined with CS, we studied thechannel estimation based on CS in underwater acoustic SISO-OFDM system andunderwater acoustic MIMO-OFDM system, regarding the underwater acoustic coherent multipath channel model based on ray theory as the object, we focus on studying theunderwater channel estimation based on Orthogonal Matching Pursuit (OMP) andCompressive Sampling Matching Pursuit (CoSaMP) algorithms and comparing with thetraditional Least Square (LS) channel estimation algorithm. Simulation results show thatcompare with LS and OMP algorithms, CoSaMP algorithm has better performance, whichcan obtain accurate CSI with fewer pilots so as to improve the frequency spectrumutilization. Moreover, the optimal number and optimal location of pilots are ascertained bysimulations and CS theory analysis in CoSaMP underwater acoustic channel estimation,providing a theoretical guiding for the practical application of underwater acousticcommunication.