基于任意麦克风阵列的声源二维DOA估计算法研究  

Speech source 2D DOA estimation algorithm based on random microphone array

作　　者：居太亮[1] 彭启琮[1] 邵怀宗[1] 林静然[1] 

机构地区：[1]电子科技大学通信与信息工程学院,四川成都610054

出　　处：《通信学报》2005年第8期129-133,共5页Journal on Communications

基　　金：华为高校科技基金资助项目(YJCB2004013MU)

摘　　要：对基于麦克风阵列的声源定位技术进行了研究,分析了基于麦克风阵列的远场信号模型,并结合子空间的方法推导出了声源二维(水平角和俯仰角)DOA估计——2D-MUSIC算法,该算法适用于任意拓扑结构的麦克风阵列。利用MATLAB仿真工具,对几种典型阵列结构进行了对比分析,提出了2种新型的三维麦克风阵列:均匀球面阵和三维均匀直线阵。仿真结果表明,提出的DOA估计算法在二维的均匀圆阵、三维的均匀球面阵和三维均匀直线阵中,均能得到较好的DOA估计效果。The speech source localization technology based on microphone array had been researched and a far-field signal model based microphone array was obtained. The two dimensions （2D, plane angle and polar angle） speech source DOA estimation algorithm had been proposed by combining this model and subspace methods. Using the MATLAB, several type of array structure was contrastively analyzed. Two kinds of new microphone army, 3D uniform spherical surface array and 3D uniform linear array were designed to 2D-DOA estimation because of ambiguity DOA of the uniform linear microphone array. Simulation results show the high performance of the DOA estimation is given when the proposed algorithm is used in planar uniform circle array, 3D uniform spherical surface array and 3D uniform linear array.

关 键 词：麦克风阵列 远场DOA估计 2D-MUSIC算法 语音信号处理 子空间方法 

分 类 号：TN911[电子电信—通信与信息系统]