长时延扩展水声信道的联合稀疏恢复估计

对具有长时延扩展的水声信道，传统的信道估计算法如最小二乘法将在大量零值抽头产生严重的估计噪声，导致估计性能下降，同时信道估计时所需的较高估计器阶数大大提高了运算复杂度。压缩感知信道估计方法可有效利用多径稀疏特性改善性能，但需采用较大的训练序列长度以保证稀疏恢复精度，由此导致额外的系统开销。利用水声信道多径稀疏结构在数据块间存在的相关性，建立基于分布式压缩感知的长时延水声信道联合稀疏模型，从而可利用同步正交匹配追踪算法进行联合重构，以进一步减小系统的训练序列开销，提高估计性能。最后通过仿真和海上实验验证了所提方法的有效性。

Jointing sparse recovery estimation algorithm of underwater acoustic channels with long time delay spread

Efficient estimation of underwater acoustic channels with a large time delay spread was addressed. For the conventional channel estimation methods such as LS, this type of channel estimation would produce serious estimation noise in zero-value taps which lead to poor performance of channel estimation. At the same time, a large time delay spread posed significant difficulties such as large channel order and the corresponding huge computation complexity. Compressed sensing (CS) channel estimation algorithm offered a solution to this problem by exploiting the sparsity of channel to improve the estimation performance. However, to ensure acceptable estimation performance, a long training sequence was needed, which unfortunately would cause additional overhead. A method was proposed which exploiting the joint correlation of sparse multipath structure between adjacent data blocks to deal with the estimation of long time delay channels under the framework of distributed compressed sensing (DCS).Thus the large time delay underwater acoustic channels can be jointly reconstructed by the simultaneous orthogonal matching pursuit (SOMP) algorithm to facilitate the system overhead reduction and estimation performance improvement. Simulation as well as the sea trial results indicate the effectiveness of the proposed method.