节点文献
矢量传感器阵列技术研究
The Study of Array Processing Technology Based on the Vector-Sensor
【作者】 吕钱浩;
【导师】 杨士莪;
【作者基本信息】 哈尔滨工程大学 , 水声工程, 2004, 博士
【摘要】 未来的新型阵列系统将具备两大特征:新的体制和新的信号处理方法。为了解决这些问题,二十世纪晚期,前苏联和美国等发达国家开发研制了将振速的测量与传统的标量测试相结合的矢量传感器阵列技术。矢量传感器由传统的无指向性标量传感器和具有与频率无关的偶极子自然指向性的质点振速传感器复合而成,它可以空间共点、同步测量声压和质点振速的各正交分量,从而得到含有声源强度信息和方位信息的声强矢量。矢量水听器比传统的声压传感器可以测量得到更全面的声场信息,从而为解决诸多的阵列探测问题提供了新的思路;这些都使得矢量传感器技术受到了各国的空前关注,成为阵列处理领域最为活跃的研究方向之一。 本文在回顾矢量传感器和矢量传感器阵列的发展历史、现状和应用的基础上,结合阵列技术最新发展趋势,围绕矢量传感器阵列的工程应用,对于矢量传感器新型体制阵列系统急需解决而尚未有公开文献表明进行研究的三个重点方向:宽带波束形成,高分辨率方位估计和阵列误差自适应校正作了较为深入的研究。 1)系统回顾了矢量传感器阵列技术的发展历程以及它们在国内外的研究应用情况,建立了矢量传感器阵列波束形成模型,首次详细探讨矢量传感器阵列波束形成方法的分类,定义了Ⅰ类子阵波束形成和Ⅱ类子阵波束形成以及混合波束形成方法,给出了常见的矢量传感器线列阵和圆阵的阵列流形并对线列阵指向性公式进行了推导,讨论了矢量传感器阵列自适应波束形成技术。 2)在矢量传感器阵列处理领域首次进行了恒定束宽波束形成器设计,提出基于自适应滤波器设计的矢量传感器阵列宽带波束形成器设计方法,该方法避免了复杂的数学推导,只需给出期望的波束图,就能由计算机完成整个设计过程,本章所提方法与传统的宽带波束形成设计方法相比,对阵列的阵型和阵元的指向性没有任何限制,不仅适用于矢量传感器阵列,也可以应用于常规标量传感器阵列,可以推广到圆阵,平面阵或者体积阵,方法具有更为广阔的应用范围。 3)首次对实际意义上的矢量传感器阵列波束域高分辨率估计技术作了研究,充分利用矢量传感器所具有的优越性,并结合波束域高分辨率估计的优点,得到了别于其他常规的基于矢量传感器的高分辨率估计方法,给出了哈尔滨__l:’-程大学博十学位论文 波束域协方差矩阵的表达式,并对其性能作了详细分析,本文方法的优越 性体现在以下几个方面:1)把声场看作是矢量场,波束是在矢量场形成; 2)矢量传感器阵列波束域高分辨率方位估计信噪比门限不仅要比阵元域 低,而且比常规阵元域高分辨率估计算法低;3)本文算法的改进无需任何 先验知识,直接进行盲波束形成,且可以大大节省计算量;4)配置灵活, 由于输出通道的繁多,可灵活配置不同形式的波束输出进行后置高分辨率 处理;本章还提出了一种改进的矢量传感器阵列波束域方位估计算法,针 对相干源信号的方位估计,本章提出了矢量传感器阵列平滑算法,有效的 得到了相干源入射的DOA估计结果。并且结合矢量传感器宽带波束形成 算法,给出了宽带信号的矢量传感器阵列波束域方位估计算法思路。4)首次对矢量传感器阵列误差的校正作了研究,建立了矢量传感器阵列误差 模型,对各种阵列误差对矢量传感器阵列的波束形成和波束域方位估计性 能的影响作了详细分析,矢量传感器阵列误差的校准要比常规阵列困难。 本文提出了矢量传感器阵列的自适应综合校准补偿方法,还提出利用单矢 量传感器速度协方差矩阵分解来求解目标源方位,使得我们不要求校准源 是合作性的,也无需知道校准源的先验知识,这比起常规辅助源校准方法 具备较大宽容性。 随着高性能阵列处理技术及数字信号处理硬件的快速发展和矢量传感器制造工艺的不断成熟,矢量传感器将在新型水雷、低频远程智能鱼雷、声纳浮标、低频拖曳线列阵和新型舰壳声纳,灵巧传感终端,战场监视系统,反隐形雷达,通讯等军民用领域取得良好的发展应用前景。矢量传感器阵列技术是构成新概念传感器阵列的核心技术,必将成为未来世界各大国竞争的焦点之一。作者希望本文有助于促进此项技术的深入研究以及进一步的工程应用和推广,由于本文一直把标量传感器看作是矢量传感器阵列的一个子阵处理,因此本文的方法同样适用于常规传感器阵列,虽然本文主要以矢量水听器为主要研究对象,但是可以很容易的推广到电磁矢量传感器(所不同的是电磁矢量传感器输出电场矢量和磁场矢量六个量)阵列,因此也可以应用于雷达,移动通讯等领域。关键词:矢量传感器,矢量传感器阵列,阵列信号处理,自适应波束形成,宽带信号处理,高分辨率方位估计,阵元误差,阵列校准,水声工程
【Abstract】 The future new type sensor array system possesses two features: fresh physics mechanism system and new signal processing methods. In order to resolve these issues , the former Soviet Union and the United States have been developed a new type sensor-vector sensor in 20th century late period , which measures the three dimensional acoustic (or electromagnetism field) particle velocity vector, as well as the acoustic(electromagnetism) pressure at one location. By preserving directional information that is present in the structure of the velocity field, they gain a number of advantages over traditional arrays of scalar sensors, such as hydrophones and microphones, so the vector sensor array technology open a new eye on resolving many problems of array processing, which attracts the all countries unprecedented concern and also become hot topic in domain of array processing.Based on the historical retrospect of the development, the actuality, and applications of the vector sensor array, the thesis, through combined with the freshest development tendency of array technology , research the key techniques home as flows: adaptive beamforming, wide band beamformer, high revolution direction estimation and self calibration using arrays of vector sensor,which is pivotal for engineering application using vector sensor array and not yet released in actual open literature.1) The development course and application situation of vector sensor array are reviewed systemically, the beamforming model using vector sensor array is established, it is the first time to classify the beamforming methods using vector sensor array in detail, Type I and II beamformer are brought forward, combined beamforming method is also put forward. The array manifold of the common circular array and linear array founded on vector sensor elements are derivated.The directiviry expression of linear vector sensor array is also showed in the thesis;2) It is the first time to design broadband beamforer in the vector sensor array processing territory, an efficient design method based on adaptive technique for wideband vector sensor array with frequency invariant beam pattern is presented, for a given beam pattern, the all design process can be completed automaticallyby computer without any complicated mathematic derivation. The proposed method can not only used in any vector sensor array system but also in any scale array system, which have no restriction on array geometry;3) This work represents the first paper relating the high revolution direction estimation methods in beam space to an array of vector sensors. This present algorithm further distinguishes itself through sufficiently utilizes the advantage that the vector sensors provide from other beam space DOA algorithms in several regards: 1) beams are formed in the vector-field domain; 2) produce at moderate signal-to-noise ratio (SNR) estimation performance comparable to the traditional scale array methods;3) these vector-field beams are formed blindly using no apriori source information and coarse estimates of the direction cosine are derived to start off MUSIC’s iterative search so it demand less computation; 4) deploy agilely , since the fact that the output channel is various .The vector sensor array space smooth algorithm is also presented here for handling coherent source signal. Moreover unite with the vector sensor broadband beam formation algorithm ,an idea which, aim at broadband signal bearing estimation using vector sensor array in beam space is also given out;4) This thesis represents the first paper relating array calibration to an array of vector sensors, the model of array errors for a vector sensor array is established, it can be find that vector sensor array errors’ calibration is more complicated and difficult than that of scale array through detailed analyses. In the thesis a novel adaptive competitive approach of array calibration for an array of vector sensors is presented which can calibrate errors of gain, directivity, phase, position and other characte