有限采样下低频声源聚焦波束形成的定位研究

Low frequency source location based on focusing beamforming under finite sample support

依赖于基阵协方差矩阵的算法,要求有足够样本容量来实现稳健的性能。针对常规聚焦波束形成在低频段分辨能力不足以及有限的数据长度,本文研究了一种基于最小方差无失真响应(MVDR)的重叠分子阵算法。仿真实验表明,重叠分子阵MVDR算法在低快拍数、低信噪比情况下,稳健性优于全阵列MVDR算法。在海试数据中,利用一条孔径为109m的光纤水听器阵列,验证了该算法可以在低快拍数下对50m内低频段100～200Hz的目标实现分辨率为0.5m的近场定位。

The motivation of this study lies in the following two observations. First, the resolution of conventional focused beamforming is limited at low frequency. Second, for estimating the sample covariance matrix to achieve stable detection performance, enough sample size is wanting. In this paper, an overlapped subarrays focused beamformer based on Minimum Variance Distortionless Response（MVDR） is researched. It is demonstrated, via simulation, that the minimum-variance, overlapped subarrays beamformer is a solid, viable approach to full array processing in small snapshots, low signal-to-noise（SNR） situations. The result of experiment in-sea shows that the location resolution can achieve the 0.5 m level within the range of 50 m on using a fiber optic hydrophone array which has an aperture of 109 m.