水声阵列探测算法及系统关键技术研究

发布时间：2019-05-05 19:13

【摘要】：水下目标的高精度、高分辨、快速定位算法及水声阵列探测系统的设计是水声阵列探测技术的核心内容。在复杂的水声环境下，水下目标的精确快速定位与跟踪，是新一代水下探测装备急需解决的关键问题。本论文针对复杂水声环境下的实际应用需求，对水下目标的定位理论与方法、对影响目标定位精度的阵列误差校正方法和阵列通道间同步采样方法、对水声阵列探测系统这一实现目标定位的硬件基础分别进行了研究。本论文的主要研究内容可归纳如下：第一，针对近远场混合源的分类与定位问题，首先，提出了一种采用交替搜寻方法进行二维混合源分类与距离(Range)参数估计的算法，它以较低的计算量实现了二维混合源的分类与距离参数估计。然后，又提出了一种采用求根比较方法进行二维混合源分类与定位的算法，它具有更低的计算量，能充分利用阵列的孔径。最后，对前述方法进行扩展，提出了一种三维混合源定位算法，实现了方位角、仰角、距离参数的高精度联合估计。第二，针对传统等距线阵孔径小而限制角度分辨率的问题，利用稀疏线阵构造了一个十字型稀疏阵列，并提出了一种能实现阵列孔径扩展的三维近场源定位算法。该算法获得了更大的阵列孔径，避免了参数配对，计算量小，能同时定位更多的信号源，具有更好的角度分辨率和参数估计精度。第三，针对单信号源定位以及阵列探测系统部分通道失效的问题，首先，基于稀疏线阵，提出了一种远场单源一维到达角(Direction-Of-Arrival, DOA)估计算法及其改进算法；改进算法具有阵列所需阵元少、角度分辨率高、角度参数估计精度高、计算量小的优点。然后，基于此一维DOA估计算法，设计了两种稀疏垂直阵列结构，并结合二维角度变换技术，实现了远场单源的二维DOA估计。第四，针对两种不同的阵列幅相误差模型，提出了两种校正原理类似的快速幅相误差有源校正算法。它们都不需进行矩阵的特征值分解，计算量非常小，同时，它们的校正精度与相位误差的大小无关，且能分别被使用到两种不同的阵列幅相误差模型中，通用性好。第五，针对传统水声阵列探测系统同步采样精度不高的问题，根据不同类型信号传输接口模块的特点，提出了两种阵列探测系统高同步精度采样方法；主要包括误差产生机理的分析，，同步采样模型的建立，传输延时的估计、测量与补偿方法的提出，克服相位抖动方法的提出、残余同步采样误差的分析等；最后，通过实验验证了所提出算法的优越性。第六，对本项目组已研制出的两套水声阵列探测系统的主要模块及成缆过程进行了简单介绍；并基于其中一套系统，通过湖试实验，对本论文中所提出的阵列幅相误差校正算法和单信号源DOA估计算法进行了实验验证，实验结果证明了这两个算法的有效性。
[Abstract]:The high precision, high resolution, fast positioning algorithm of underwater targets and the design of underwater acoustic array detection system are the core contents of underwater acoustic array detection technology. In the complex underwater acoustic environment, the precise and rapid positioning and tracking of underwater targets is a key problem urgently to be solved in the new generation of underwater detection equipment. According to the practical application requirements of complex underwater acoustic environment, the theory and method of underwater target positioning, the method of array error correction and the synchronous sampling method between array channels, which affect the accuracy of target location, are presented in this paper. The hardware foundation of underwater acoustic array detection system is studied. The main contents of this thesis can be summarized as follows: first, aiming at the classification and location of near-and far-field mixed sources, firstly, An alternative search algorithm for two-dimensional hybrid source classification and distance (Range) parameter estimation is proposed in this paper. The two-dimensional hybrid source classification and distance parameter estimation are implemented with low computational complexity. Then, an algorithm for two-dimensional hybrid source classification and location using root-finding comparison method is proposed, which has lower computational complexity and can make full use of the aperture of the array. Finally, a three-dimensional hybrid source localization algorithm is proposed to achieve high-precision joint estimation of azimuth, elevation and distance parameters. Secondly, aiming at the problem that the aperture of the traditional isometric linear array is small and the angular resolution is limited, a cross-shaped sparse array is constructed by using the sparse linear array, and a three-dimensional near-field source localization algorithm which can realize the expansion of the aperture of the array is proposed. The algorithm can obtain larger array aperture, avoid parameter pairing, reduce computational complexity, locate more signal sources at the same time, and have better angle resolution and parameter estimation accuracy. Thirdly, to solve the problems of single signal source location and partial channel failure of array detection system, firstly, based on sparse linear array, a far-field single source one-dimensional arrival angle (Direction-Of-Arrival, DOA) estimation algorithm and its improved algorithm are proposed. The improved algorithm has the advantages of fewer array elements, high angle resolution, high accuracy of angle parameter estimation and less computation. Then, based on this one-dimensional DOA estimation algorithm, two kinds of sparse vertical array structures are designed, and the two-dimensional DOA estimation of the far-field single source is realized by combining the two-dimensional angle transformation technique. Fourth, for two different array amplitude and phase error models, two fast active correction algorithms for amplitude and phase error are proposed, which are similar in principle. They do not need to decompose the eigenvalues of the matrix, and the calculation is very small. At the same time, their correction accuracy is independent of the size of the phase error, and can be used in two different array amplitude and phase error models. Fifthly, aiming at the problem that the precision of synchronous sampling is not high in the traditional underwater acoustic array detection system, according to the characteristics of different types of signal transmission interface modules, two high-precision sampling methods of array detection system are proposed. It mainly includes the analysis of the mechanism of error generation, the establishment of synchronous sampling model, the estimation of transmission delay, the proposed method of measurement and compensation, the method of overcoming phase jitter, the analysis of residual synchronous sampling error, and so on. Finally, the superiority of the proposed algorithm is verified by experiments. Sixth, the main modules and cable-forming process of two sets of underwater acoustic array detection system which have been developed by the project team are briefly introduced. Based on one of the systems, the calibration algorithm of array amplitude and phase error and the DOA estimation algorithm of single signal source are verified by lake experiment. The experimental results show that the two algorithms are effective.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TB566

【引证文献】