一种用于分布式阵列的球谐波域声源定位方法

Sound Source Localization Using Distributed Microphone in Spherical Harmonics Domain

本文提出一种适用于任意阵型和阵元指向性的球谐波域声源定位方法,能够在较宽的频域范围内,尤其是低频,提供较高的空间分辨率。水下噪声源的高分辨识别具有重大意义。传统阵列信号处理方法对低频噪声源的精确定位要求阵列具有较大孔径和较多的阵元数,导致系统过于庞大且成本较高。我们基于声场的球谐波表达和变换,采用分布在一定空间区域内的多个阵列估计该区域的球谐波系数矢量,对系数矢量进行信号处理实现声源定位。理论证明了该方法具有理想的空间选择特性。在一种特定的阵元分布下,仿真研究了该算法的方位谱估计性能以及阵元不一致性和位置误差对声源定位性能的影响。仿真结果显示,该算法在低频具有较高的空间分辨率且误差对算法定位性能的影响有限。

A sound source localization method in spherical harmonics domain for arbitrary directivity of distributed sensors is proposed, which can provide high resolution in a wide frequency range. Based on the spherical harmonic representation of a sound field and its translation, we estimate the spherical harmonic coefficients by using multiple sensors distributed in a certain region. Sound source localization can be realized by processing these coefficients. Theoretical analysis shows the ideal spatial selection characteristics of the method. Simulations are performed under a specific sensor distribution, which show that the method has high resolution at low frequency and is robust to the error of the position of array elements.