基于麦克风阵列的声源定位算法及系统实现

发布时间：2018-09-18 12:35

【摘要】：本文的研究内容是国家自然科学基金资助项目— “基于轮轨噪声的轨道安全检测原理研究”的一部分。项目试图对列车轮轨相互作用所产生噪声信号进行研究,目的是提取噪声所包含的轮轨系统的故障信息,能够对铁路轨道系统的故障状况进行实时的动态检测,以便于优化我国铁路运输安全的故障检测能力。本文是该项目的一个基础环节,也是该项目的重要组成部分,目的是通过对轮轨噪声源位置的确定,为下一步进行轮轨缺陷识别提供有效的位置信息。本文主要内容如下所述:第一,构建数据采集实验平台。数据采集平台是实验的硬件部分,其主要任务是采集声音信号并进行简单信号调理,并且实现对复杂音频信号的采集与处理。第二,针对此项目具体要求特性,对声音的近场和远场模型进行判定,判明本文研究的声场类型及定位范围:介绍国内外典型的三种声源定位算法,根据其特点选择适合本文的算法深入研究。第三,研究基于可控波束的定位法,进行仿真分析然后优化算法;将缩减空间搜索的算法引入到轮轨噪声故障诊断定位领域,实验结果表明改进后算法精度提高36%,算法耗时仅为之前28%,同时该算法具有良好的抗噪性。第四,研究基于时延的定位法,仿真分析并建立一套实际的定位系统进行实验,实现了等边三角形阵列对三维声源坐标的定位功能,结合本项目的情况进行实验分析,发现实验结果得最大误差值在误差允许范围之内,方案可行。
[Abstract]:The research content of this paper is a part of the project supported by the National Natural Science Foundation-the Research on the principle of track Safety Detection based on Wheel-Rail noise. The purpose of the project is to extract the fault information of the wheel-rail system contained in the noise, and to detect the fault state of the railway track system in real time. In order to optimize the fault detection ability of railway transportation safety in China. This paper is a basic part of the project and an important part of the project. The purpose of this paper is to provide effective location information for the identification of wheel-rail defects by determining the location of wheel-rail noise sources. The main contents of this paper are as follows: first, build a data acquisition experimental platform. The data acquisition platform is the hardware part of the experiment. Its main task is to collect the sound signal and adjust the simple signal, and realize the acquisition and processing of the complex audio signal. Secondly, according to the specific requirements of this project, the near-field and far-field models of sound are judged, and the types and localization range of sound field studied in this paper are determined. Three typical sound source localization algorithms at home and abroad are introduced. According to its characteristics, the algorithm suitable for this paper is selected for further study. Thirdly, the localization method based on controllable beam is studied, the simulation analysis is carried out and then the algorithm is optimized, and the reduced space search algorithm is introduced into the fault diagnosis and location field of wheel-rail noise. The experimental results show that the accuracy of the improved algorithm is improved by 36%, the time consuming of the algorithm is only 28%, and the algorithm has good anti-noise performance. Fourthly, the localization method based on time delay is studied, and a set of actual positioning system is established to carry out experiments. The function of equilateral triangle array to locate 3D sound source coordinates is realized. The experiment analysis is carried out according to the situation of this project. It is found that the maximum error of the experimental results is within the range of error, and the scheme is feasible.
【学位授予单位】：西安理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN912.3

【参考文献】