时变水声信道的动态压缩感知估计

水声多径信道具有的稀疏特性已被应用于设计稀疏估计方法提高信道估计性能。然而,水声信道具有的快速时变特性,给传统稀疏信道估计方法带来了很大的困难。考虑到水声信道除了高时变性多径,还存在着相对静止,缓慢变化的直达径或者海底反射径,通过将水声信道建模为由静态和时变稀疏支撑集组成,把时变水声信道估计转化为动态压缩感知问题。结合卡尔曼滤波和压缩感知理论,并采用原始对偶追踪算法求解Dantzig selector模型,从而实现对复数域基于卡尔曼滤波器的压缩感知稀疏求解问题的处理。信道时变条件下的数值仿真及基于信道估计的判决反馈均衡器的海上实验结果表明,该算法相对经典的正交匹配追踪和最小二乘QR分解算法具有较明显的性能改善。从而说明,通过对时变水声信道进行动态压缩感知估计可有效提高估计性能。

Dynamic compressed sensing estimation of time varying underwater acoustic channel

It has been recognized that, while the multipath structure offer the potential for sparsity exploitation by the means of compressed sensing, the rapidly time varying arrivals pose significant difficulties to the UnderWater Acoustic(UWA) channel estimation. Except the highly time varying arrivals caused by dynamic surface, generally there exist relatively stationary or slowly changing arrivals caused by direct path or bottom. By modeling the time varying UWA channels as sparse set consisting with constant and time-varying supports, the estimation of time varying UWA channel is transformed into a problem of Dynamic Compressed Sensing(DCS) sparse recovery. Via the combination of Kalman Filter(KF) and Compressed Sensing(CS), the Primal Dual Pursuit(PD-Pursuit) for the Dantzig Selector(DS) is adopted to pursue the complex domain solution of UWA channel estimation. Numerical simulations demonstrate the superiority of the proposed algorithm. In the form of a Channel-Estimation-based Decision-Feedback Equalizer(CE-DFE), the experimental results with the field data obtained in a shallow water acoustic communication experiment indicate that,the proposed algorithm outperforms the classic Least Square QR-factorization(LSQR) or Orthogonal Matching Pursuit(OMP) algorithms. Therefore, it is shown that the DCS estimation can improve the estimation performance effectively in time varying UWA channel.