基于特征分解的一步最小二乘声源定位算法  

A Closed-form One-Step Least-Square Algorithm of Acoustic Source Localization Based on Eigenvalue Decomposition

作　　者：张莉[1] 吕明[1] 

机构地区：[1]电子科技大学电子工程学院,四川成都610054

出　　处：《电声技术》2007年第5期42-45,共4页Audio Engineering

摘　　要：针对基于传声器阵列的声源定位闭式解问题,提出了一种基于特征值分解的闭式一步最小二乘定位算法,并通过计算机仿真比较了其与球形插值法及球形内插法的性能,表明球形插值法和一步最小二乘法具有等效的统计性能并且计算更为简便,且二者性能优于球形内插。To solve the problem of closed-form solution of acoustic source localization based on microphone army, a closed-form OSLS（One-Step Least-Square） algorithm based on eigenvalue decomposition is proposed. The statistical performances of the OSLS algorithm, the SI（Spherical Interpolation） method and the SX（Spherical Intersection） method are compared via computer simulation. It is shown that the OSLS algorithm is statistically equivalent to the SI method but with less computational complexity, which appears superior to the SX method.

关 键 词：传声器阵列 声源定位 闭式解 特征值分解 

分 类 号：TN912.3[电子电信—通信与信息系统]