题名多宽带源波达方向估计稳健算法研究
作者滕鹏晓
学位类别博士
答辩日期2007-06-08
授予单位中国科学院声学研究所
授予地点声学研究所
关键词波达方向估计 宽带源 多源 稳健 高分辨 多阵列
其他题名Research on robust algorithm of direction of arrival estimation for multiple wideband sources
学位专业信号与信息处理
中文摘要波达方向估计是阵列信号处理中一个非常重要的研究领域,由于其在雷达、声呐、海洋测绘、医学成像、数字移动通信及机器故障诊断等许多领域中有着广泛且重要的应用背景而受到重视。近些年来,稳健的波达方向估计算法一直是研究的热点,因为这决定着算法能否在工程中得到实际的应用。空间分辨率也是重要的性能参数,它预示着估计算法分辨空间相近信源的能力。另外,随着宽带信号在各个领域的广泛应用以及空间通常会同时存在着多个信源,研究能够同时估计出多个宽带信号源的波达方向必然成为一项非常重要的研究内容。 本文就不同的应用背景和实际环境,围绕波达方向估计的稳健性、多源估计、宽带源估计及高分辨估计等方面做了诸多研究工作,主要工作和创新点概括如下: (1) 针对课题背景中的目标信号具有较强的线谱成份,提出一种基于谐波提取的宽带源波达方向估计算法,并分析、结合了相干信号子空间和非相干信号子空间算法的优点。仿真结果表明估计性能得到明显的提高。 (2) 为了抑制实际环境中有色噪声的影响,提高算法对噪声的稳健性,提出了一种能够应用于空间相关色噪声环境下的相干宽带源波达方向估计算法。在通常情况下,相干信号子空间算法假定了噪声信号为空时白噪声,但是在实际应用中,尤其当阵列的孔径相对较小时,噪声信号在很多情况下为有色噪声,这导致了相干信号子空间类算法聚焦误差的增加和波达方向估计性能的严重下降。为此,提出了一种基于差分矩阵的宽带聚焦变换算法,该算法有效抑制了有色噪声,从而减小了聚焦误差。仿真和实验结果表明,提出的方法能适用于低信噪比环境下的相干宽带源的波达方向估计,稳健性及分辨率得到明显提高。 (3) 研究多传声器阵列联合处理算法,旨在提高定位的精度和在复杂混响环境下定位的稳健性。由于受到环境的干扰,目标信号非平稳性的影响及受传声器阵列物理尺寸的限制,通常单一传声器阵列对信号源定位,不能得到较高的精度,而在空间分布的多个传声器阵列的联合处理则提供了一种精确定位的解决方案。提出的算法通过把多个传声器阵列接收的数据协方差矩阵变换到同一个参考阵列上,然后再进行声源定位,这样充分利用多个传声器阵列的数据,抑制噪声,提高估计的精度和稳健性。通过数据仿真和实验来检验提出的算法的性能,仿真和实验结果表明定位精度和稳健性得到提高;在实验中,同时检验了算法对移动声源的定位性能。由于真实房间内混响的存在和以及声源的运动引起信号相关性的下降,导致定位结果方差较大,最后使用粒子滤波算法对定位结果进行后处理,最终得到较好的定位结果。
英文摘要The direction-of-arrival (DOA) estimation is an important research area of array signal processing, with extensive applications in many fields such as Radar, Sonar, medical imaging , digital mobile communication and machinery trouble shooting and diagnosis. In recent years, more attention has been paid to the robustness of DOA estimation algorithm, which is the key factor in engineering applications. The spatial resolution is also an important performance indicator, which implies the capability of the separation of the closely spaced sources. In addition, to deal with the growing applications of the wideband signals in many fields and the practical environments where multiple sources exist, it becomes more important to investigate the DOA estimation algorithm which can separate the multiple wideband sources. This dissertation aims at the development of the high-resolution and robust estimation for multiple wideband sources with consideration of different application background and practical acoustic environment. The main contributions of the research work can be summarized as follows: (1) A type of signals which involve strong line spectra is considered. Under this condition, a DOA estimation method is introduced based on the extraction of harmonics contained in the signals. In virtue of both the coherent signal subspace method and non-coherent signal subspace method, the DOA estimation results are improved effectively. (2) A robust DOA estimation algorithm for coherent wideband sources in colored noise fields is investigated. Generally, the noise is unknown and correlated from sensor to sensor in practical applications, especially for arrays with relative small apertures. The spatially correlated noise incurs an increase in focusing error and a severe degradation in the DOA estimation. To deal with the problem, a new method of focusing transformation based on differentiating covariance matrix is proposed to eliminate noise, and hence reduce the focusing error. The simulation and experiment results demonstrate the effectiveness of the proposed method. (3) A new DOA estimation method based on the collaborative processing of multiple arrays is proposed to improve the robustness and accuracy of localization. While the accuracy of source localization using a single array is low, the collaborative processing using multiple arrays distributed in space provides an effective solution to the problem. The covariance matrices of multiple arrays are coherently transformed into that of a reference array, and locations of multiple sources are estimated thereafter. With the method, the resolution threshold and error are reduced, and the robustness and accuracy of localization are improved significantly. The simulation and the experiment are carried out to demonstrate the effectiveness of the proposed algorithm. To reduce the variance of localization for a moving source in a reverberant room, the particle filter is further used to improve the localization accuracy.
语种中文
公开日期2011-05-07
页码113
内容类型学位论文
源URL[http://159.226.59.140/handle/311008/156]  
专题声学研究所_声学所博硕士学位论文_1981-2009博硕士学位论文
推荐引用方式
GB/T 7714
滕鹏晓. 多宽带源波达方向估计稳健算法研究[D]. 声学研究所. 中国科学院声学研究所. 2007.
个性服务
查看访问统计
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。


©版权所有 ©2017 CSpace - Powered by CSpace