中文摘要：

为了实现稀疏水声信道的快速高效估计,提出一种新的基于有效抽头的信道参数模型降低运算复杂性,并采用进化算法以最小二乘误差为适应度函数对信道响应中有效抽头的阶数、位置和权重同时进行进化操作寻优,避免了有色输入造成的相邻抽头耦合导致的性能下降。仿真实验结果表明：与传统LMS算法及基于有效抽头检测的稀疏信道估计方法相比,该算法在有色输入信号下具有更优越的估计性能和收敛速度,为稀疏水声信道处理提供了一类新方法。

英文摘要：

To estimate sparse underwater acoustic channels, a filter structure only containing positions and values of active taps is proposed to reduce computational complexity compared with traditional transversal FIR filters. Consider the coupling effects of active and inactive taps induced by colored input which leads to failure of detection guided algorithm, the order, as well as the positions and values of the active taps are evolved simultaneously with least squares error as fitness function. Simulation results demonstrate that an active taps and evolutionary method show better efficiency for sparse channel estimation in case of colored inputs.