麦克风阵列声源定位系统的阵列优化设计研究

发布时间：2018-11-01 14:36

【摘要】：随着阵列信号处理与语音处理在雷达、声纳及无线通信等众多军事及国民经济领域应用中的不断发展，阵列拓扑结构与声源二维DOA估计等方面的研究，尤其是前者对后者的性能影响越来越受到人们的重视。根据阵列信号处理的理论，阵列拓扑结构对阵列处理系统的性能具有重要影响。因此，以二维DOA估计为切入点，开展阵列优化拓扑结构的研究具有重要的科学意义和应用价值。在综述国内外相关阵列拓扑结构及其相关DOA估计技术研究进展的基础上，论文首先给出了声源定位中阵列拓扑研究的基础理论和基础实验分析；系统研究了均匀线阵、嵌套线阵、均匀L型阵、矩形阵、圆阵和同心圆阵等不同拓扑结构及其阵列方向向量；详细分析了MUSIC、Capon和ESPRIT等经典的二维DOA估计方法；给出了相应的阵列波束图和空间谱图的仿真结果。其次，论文介绍了阵列优化设计中的优化算法的理论基础，系统阐述了遗传算法的基本原理和实际中关键参数的优化选择。考虑到阵列拓扑结构的优化方法可以分为两种：一种是优化各个阵元位置，进而校正阵元坐标；另一种是优化阵元个数，保留的阵元位置不变，论文利用遗传算法对阵列拓扑结构中阵元个数进行优化，通过构造由二维MUSIC空间谱函数欧式距离和优化后阵元个数共为变量的适应度函数，声源2-D DOA估计精度为停止条件，采用经典的二维MUSIC算法，对均匀L型阵、均匀矩形阵、均匀圆形阵和均匀同心圆阵等麦克风阵列进行了优化仿真设计。仿真实验结果表明，与其他阵列相比，论文方法所得到的优化阵列拓扑结构，能够较准确的估计出声源的来波方向，且能节省一定数目的阵元。
[Abstract]:With the development of array signal processing and speech processing in many fields of military and national economy, such as radar, sonar and wireless communication, the research of array topology and two-dimensional DOA estimation of sound source, etc. Especially, the influence of the former on the performance of the latter has been paid more and more attention. According to the theory of array signal processing, array topology has an important effect on the performance of array processing system. Therefore, the research of array optimization topology based on two-dimensional DOA estimation has important scientific significance and application value. On the basis of summarizing the research progress of related array topology structure and related DOA estimation technology at home and abroad, the basic theory and experimental analysis of array topology research in acoustic source location are presented in this paper. Different topologies such as uniform linear array, nested linear array, uniform L-type array, rectangular array, circular array and concentric circular array and their array direction vectors are systematically studied, and classical two-dimensional DOA estimation methods such as MUSIC,Capon and ESPRIT are analyzed in detail. The simulation results of the array beam map and the spatial spectrum diagram are given. Secondly, the paper introduces the theoretical basis of the optimization algorithm in array optimization design, and expounds the basic principle of genetic algorithm and the optimization of key parameters in practice. Considering the array topology optimization methods can be divided into two types: one is to optimize the location of each element and then to correct the coordinates of array elements; The other is to optimize the number of array elements and keep the position of the array elements unchanged. The genetic algorithm is used to optimize the number of elements in the array topology. By constructing the fitness function from the Euclidean distance of two-dimensional MUSIC space spectrum function and the number of optimized array elements as variables, and the accuracy of 2-D DOA estimation of sound source as stopping condition, the classical two-dimensional MUSIC algorithm is used for uniform L-type matrix and uniform rectangular array. The microphone arrays such as uniform circular array and uniform concentric circular array are optimized and designed. The simulation results show that compared with other arrays, the optimized array topology obtained by the proposed method can accurately estimate the direction of the acoustic source and save a certain number of array elements.
【学位授予单位】：辽宁工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN912.3

【参考文献】