一种改进的WSF算法在单矢量水听器多目标方位估计中的应用

Multi-source DOA estimation using a modified WSF algorithm in a single vector hydrophone applications

文章针对单个矢量水听器的多目标方位估计,提出了一种基于加权子空间拟合(WSF)的算法,该算法首先对单个矢量水听器接收数据作一任意的时间延迟,而后仿照ESPRIT算法的思路求解阵列响应矩阵,从中抽取各目标的波达方位。该算法在保留WSF算法分辨力高、估计方差小的优点的同时,通过结合ESPRIT算法的思想,克服了WSF算法计算量大,需迭代求解的缺点。由于该算法和声源频率无关,因而可直接应用于宽带声源的测向,并避免了传统ESPRIT算法中因估计延时相位而导致的频率-方位模糊问题。文中通过数值仿真和推导Cramer-Rao下界,给出了该算法的性能评价,数值仿真和湖试实验结果也充分验证了该算法的有效性。

This paper develops a new algorithm based on WSF with respect to estimate the directions of targets in the multi-source scenario using a single vector hydrophone. This method exploits the sources＇ DOA information embedded in the sources＇ velocity field by an ESPRIT algorithm and reduces the DOA variances via weighted subspace fitting techniques. It has high resolution as other eigenstructure-based algorithms, moderate computational costs, suffers no frequ- ency-DOA ambiguity, and can be directly applied to wideband source scenarios. The Cramer-Rao lower bound of mean-square-angular-error is deduced to evaluate its performance .Both numerical simulation and lake test results verify the efficacy of this proposed scheme.