声场分解的均匀圆阵实值MUSIC算法

Unitary MUSIC algorithm based on sound field decomposition for UCA

经典方位角估计算法中未考虑阵列安装支架对阵列接收信号的影响,实际中阵列安装支架必然会对阵列接收信号产生影响。以环绕在刚性球上的均匀圆阵为阵列模型,在对声场特性的分析中将声学原理和阵列信号处理技术相结合,探讨了存在刚性球形障碍物时的声源方位角估计问题。首先从声学理论出发,分析了刚性球体散射声场及声场分解,讨论了刚性球体对圆阵响应的影响;进而结合阵列信号处理技术,在对声场分解所得到的特征波束空间,利用实值MUSIC算法实现了声源方位角估计。计算机仿真表明,该算法能较好地估计出空间多个声源的方位角,且计算量小,估计精度高,具有解相关声源的能力。

Classical array signal processing algorithms used in acoustic source localization systems have been successfully used for direction of arrival （DOA） estimation in free field. In practices, array elements are mounted on some stands, which may influence received signals. Using the array model of a uniform circular microphone array （UCA） mounted on a rigid spherical baffle, the problem of direction of arrival of acoustic sources with rigid spherical baffles is discussed based on classical array signal processing and classical acoustics theory. The scattering sound field and its decomposition on the baffle are analyzed in detail based on the principles of wave propagation and scattering theory. The influence of the baffle on the UCA response is analyzed simultaneously. In the eigen-beam space derived from the decomposed sound field in spherical coordinates, bearing estimation of acoustic sources is realized by using a novel formulation of MUSIC algorithm. Computer simulations show that the algorithm has the ability of estimating the bearings of multi-acoustic sources and processing coherent acoustic sources with low computational load and high performance.