用MUSIC法确定声源频率和方位时误差的校正方法  

A Calibration Method of Model Error for Frequencies and Bearings Estimation of Sources with MUSIC

作　　者：俞靖[1] 鲍明[1] 

机构地区：[1]南京航空航天大学振动工程研究所

出　　处：《南京航空航天大学学报》1998年第6期680-684,共5页Journal of Nanjing University of Aeronautics & Astronautics

摘　　要：为确定多个平面波窄带声源的频率和方位，本文在阵列信号处理理论的基础上，提出用均匀线阵接收信号数据，并采用多信号分类法（ＭＵＳＩＣ法）对其进行频率、方位参数的分离估计，利用其正交性的谱估计来判断声源的频率和方位。在此基础上，针对实际阵列中存在的阵元位置扰动、通道特性不一致和阵元之间的耦合这三种模型误差，提出一种基于辅助源的校正方法，并得模型误差不影响声源的频率，只影响声源方位的判断。声学实验证实，通过校正得到的空间谱准确地分辨了声源的方位角，与未经校正的相比，本文提出的方法能有效地消除模型误差引起的测向误差，大大提高测向分辨率，由此证实了这种方法的有效性。Based on the theory of array signal processing, a separable approach with a method of multiple signal classification (MUSIC) and orthogonality of spectrum is proposed to estimate frequencies and azimuth angles of multiple planar wavefront noise sources with narrow band using a uniform linear array. Based on it, a method of model error calibration by additional source is proposed to reduce the three kinds of model errors in a practice, and model errors have not influence on frequencies but influence on azimuth angles. The theory in the paper is verified by acoustics experiments in anechoic room. Theexperimental results show that the method has high resolution. It shows that the method is effective.

关 键 词：线列阵 频率 声源 方位角 误差校准 多信号分类 

分 类 号：TB52[理学—物理] TN911.7[理学—声学]