The Research Of Channel Estimation Based On Structured Sparsity In Underwater Acoustic

Underwater acoustic channel is a space-time-frequency varying channel,limited channel bandwidth,and severe multipath interference,which is by far the most complicated wireless channel.Orthogonal frequency division multiplexing(OFDM)technique can use the limited bandwidth of underwater acoustic channels to effectively resist the intersymbol interference caused by multipath.However,with the increase of communication distance,the channel state is deteriorated,and the received signal is also deteriorated in amplitude,phase and frequency.Therefore,it is necessary to use channel estimation technology to track the channel state and compensate the received signal.This paper mainly studies the underwater acoustic channel estimation based on OFDM system.In the traditional underwater acoustic channel estimation,the channel is assumed as rich multipath channel.Use Nyquist sampling theorem to sample channel impulse response,which need many pilots to estimate channel.At the same time,the system information transmission rate and spectrum efficiency are reduced.Compressed sensing(CS)theory broke the traditional Nyquist sampling theorem,sampling and compressing the signal at the same time.It can highly reconstruct the original signal with a small amount of sample values.Channel estimation mainly based on sparse underwater channel model.Much ocean experiment shows that heterogeneity of seabed media leads to the acoustic ray spread as block.The underwater acoustic channel exhibits as block sparse structure.Therefore,underwater acoustic channel estimation problem can be converted into a block structure sparse signal reconstruction problem.Use structure CS theory to block sparse sample and reconstructe the channel.It can sharply reduce the number of pilots needed to reconstructe the channel,and improve the utilization rate of system frequency band.It can avoid estimating the useless zero tap,and improve the efficiency of reconstruction.This article introduce about the physical characteristics of underwater acoustic channel and the basic principle of OFDM system.Based on the theory of ray acoustic,we can set up the coherent multipath channel model.Introduce the traditional least square channel estimation method simply.Then,the compress perception theory is introduced.Based on the theory,introduce the concept of block structure sparse signal and its related reconfiguration algorithm.Mainly introduce the classic block orthogonal Matching Pursuit(Block Orthogonal Matching Pursuit,BOMP)algorithm.Based on the several orthogonal matching thought,apply the improved algorithm in underwater acoustic channel estimation.It is simple and easy to estimate the channel with pilots,and this method has been widely used.This article estimate the channel based on the pilots.Then the traditional LS algorithm,Orthogonal Matching Pursuit based on sparse model(Orthogonal Matching Pursuit,OMP)algorithm,the BOMP based on block sparse model and its improved algorithm are applied in the underwater acoustic,to reconstruct the underwater acoustic channel.On the software platform,simulate the underwater acoustic channel estimation by experiments.Simulation results show that: the BOMP algorithm,which based on block structure sparse model,is superior to the traditional LS and OMP algorithm.With the small amount of pilots,BOMP algorithm can get better performance of channel estimation and reduce the channel reconstruction time.The improved BOMP algorithm can further reduce the channel reconstruction time,with the guarantee of reconstruction precision of the BOMP algorithm.