宽带信号DOA估计算法与区域定位方法的研究及应用

发布时间：2018-06-20 16:03

本文选题：波达方向估计 + 小波变换　；参考：《安徽大学》2017年硕士论文

【摘要】：近几十年,随着阵列信号处理技术的发展,传感器成本的降低及其处理器运算速度的加快,声阵列传感器网络被广泛的应用于生活、医疗、军事等各种领域。声阵列被动探测系统利用阵列的麦克风阵元接收目标声源信号,采用阵列信号处理算法模型处理信号,实现对声源目标的探测、测向与定位。本文主要内容包括介绍典型的宽带信号波达方向估计算法,在此基础上提出基于小波变换的高分辨率测向算法,并设计出一种区域定位系统,文章通过仿真实验及应用实验验证了该系统的有效性。具体如下:首先,介绍了宽带阵列信号模型,以及基于此模型的波束形成算法、多重信号分类算法及宽带聚焦矩阵算法,给出算法原理和实现步骤,并利用仿真实验对各算法在不同信噪比、快拍数等影响因素下的误差性能进行分析。其一次,讲述了小波变换的理论基础,并将小波变换用于建立时频阵列信号模型,该模型不同于以往宽带阵列信号模型中均匀分割频带的方法,它利用小波变换的多分辨特性将接收信号的宽频带进行多分辨划分,能提取信号中更全面的时频信息。随后提出基于此模型的高分辨率测向算法,推导出该测向算法的克拉美罗界,并构造仿真实验分析算法的误差性能以及多分辨性能,验证该算法的稳健性和高分辨率特性。再次,提出一种区域定位系统的设计方案,系统主要实现对声源目标信号的检测、定位功能,方案首先对接收信号进行检测,检测是否存在目标信号,若检测出信号则提取它的频谱特征并对目标源进行定位,针对近场和远场两种情况分别提出定位几何模型和区域谱估计模型,通过搜索区域谱峰值获得目标声源定位结果。为了在保证区域谱分辨率的同时降低计算成本,先采用低分辨率谱估计得到粗略的定位估计,再对估计位置所在的小范围区域内进行高分辨率谱估计并实现最终的定位。最后,本文利用课题组现有的阵列设备,设计模拟爆炸声定位的仿真实验并且在半消声实验室进行了鸣笛声定位实验,实验验证了该定位系统的实用性。
[Abstract]:In recent decades, with the development of array signal processing technology, the reduction of sensor cost and the acceleration of processor speed, acoustic array sensor network has been widely used in life, medical treatment, military and other fields. The acoustic array passive detection system receives the target sound source signal by the microphone array element and uses the array signal processing algorithm model to process the signal to realize the detection direction and localization of the sound source target. The main contents of this paper include the introduction of typical DOA estimation algorithms for wideband signals. On this basis, a high resolution direction finding algorithm based on wavelet transform is proposed, and a region location system is designed. The effectiveness of the system is verified by simulation and application experiments. The details are as follows: firstly, the wideband array signal model, the beamforming algorithm based on this model, the multi-signal classification algorithm and the wideband focusing matrix algorithm are introduced. Simulation experiments are used to analyze the error performance of each algorithm under different influence factors such as signal-to-noise ratio (SNR) and number of beats. In this paper, the theoretical basis of wavelet transform is described, and the wavelet transform is used to establish the time-frequency array signal model, which is different from the method of evenly dividing the frequency band in the previous wideband array signal model. It uses the multi-resolution characteristic of wavelet transform to divide the received signal in a wide frequency band and can extract more comprehensive time-frequency information from the signal. Then, a high-resolution direction-finding algorithm based on this model is proposed, and the Crameiro bound of the direction-finding algorithm is derived, and the error performance and multi-resolution performance of the algorithm are constructed to verify the robustness and high-resolution characteristics of the algorithm. Thirdly, a design scheme of regional positioning system is proposed. The system mainly realizes the detection and location function of the target signal of the sound source. The scheme first detects the received signal and detects the existence of the target signal. If the signal is detected, its spectrum feature is extracted and the target source is located. The localization geometric model and the region spectrum estimation model are proposed for the near field and far field respectively. The target acoustic source location results are obtained by searching the peak value of the region spectrum. In order to ensure the spectral resolution of the region and reduce the computational cost, the low resolution spectral estimation is first used to obtain the rough location estimation, and then the high-resolution spectral estimation is carried out in the small region where the estimated position is located and the final location is realized. Finally, using the existing array equipment of the research group, the simulation experiment of simulating explosive sound location is designed, and the sound localization experiment is carried out in the semi-anechoic laboratory, which verifies the practicability of the positioning system.
【学位授予单位】：安徽大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.7

【参考文献】