一种基于泛函波束成形的传声器阵列优化方法

An approach for microphone array geometry optimization based on functional beamforming

本文基于泛函波束成形算法,提出了一种可以用于二三维声源识别的传声器阵列结构优化方法和相应的优化阵列形式。本文的渐开线螺旋臂阵列设计包含4个主要几何参数:传声器数目、阵列孔径、螺旋臂的数目和基圆的大小。通过基于改进的遗传算法的优化,得出两种优化阵列:优化渐开线螺旋阵列和优化随机阵列,目的是减小波束模式的主瓣宽度以提高传声器阵列的分辨率性能。本文的优化过程包括对最优几何参数组合的寻找和对传声器坐标直接优化两个部分,优化关注点在于中低频声源识别的分辨率性能,在对比研究中发现,上述两种优化阵列在二三维声源识别方面都表现出较好的性能。

In this paper,an optimization method of microphone array structure and the corresponding optimized array form are proposed for the identification of two and three dimensional sound sources. The design of involute spiral arm array in this paper contains four main geometric parameters:the number of microphones,the aperture of the array,the number of spiral arms and the size of the base circle. Through the optimization process based on the improved genetic algorithm,two kinds of optimization arrays are obtained:optimization of involute spiral array and optimization of random array. The purpose is to reduce the width of the main lobe of the beam mode and improve the resolution performance of the microphone array. The optimization process in this paper includes two parts:searching for the optimal combination of geometric parameters and directly optimizing the microphone coordinates. The optimization focuses on the resolution performance of low and medium frequency sound source identification. In the comparative study,it is found that the above two optimized arrays show good performance in the identification of two and three dimensional sound sources.