声矢量阵校正及测向技术研究

发布时间：2018-11-10 17:39

【摘要】：与声压阵相比,声矢量阵克服了左右舷模糊问题,扩展了基阵孔径,降低了干扰背景和检测阈,在线阵、平面阵或小阵元数的情况下优势明显。基于以上几点,声矢量阵信号处理技术成为水声领域的热点之一,美国和俄罗斯大力研究如何将声矢量阵应用于海军装备,用于水下目标的探测和定位,声矢量阵对于现代海军装备的重要性可见一斑。本文主要研究声矢量阵的校正及测向技术。众多声矢量阵测向技术的高分辨性能都是在理想阵列流型的前提下得到的,在实际情况下,声矢量阵往往存在多种阵列误差,这些阵列误差将导致声矢量阵高分辨测向算法性能严重下降,甚至失效。因此,在使用声矢量阵高分辨测向技术之前,对声矢量阵进行阵列校正是至关重要的。对于声矢量阵测向问题,现有的大部分声矢量阵方位估计算法都具有较大的运算量,这大大增加了系统的成本和负担,成为某些系统工程实现的瓶颈。另外,在水声环境中,往往存在相干源,研究声矢量阵相干源多目标分辨对声矢量阵测向技术具有重要的意义。本文的主要研究内容有:1.针对声矢量阵幅相误差及阵元位置误差有源校正问题,基于多级维纳滤波器(MSWF),提出了快速阵列误差校正的 SMSWF(Simplified Multi-Stage Wiener Filter)算法。SMSWF算法同时利用校正源的方位和波形信息对阵列误差参数进行估计,无需协方差矩阵计算和特征值分解过程,大大地减小计算量,且具有与特征分解法相同的阵列误差参数估计性能。研究发现,当单个信源入射到阵列且信源波形已知时,SMSWF算法获得的信号子空间等价于特征分解法估计的信号子空间,此时SMSWF算法能够替代特征分解法,从而可以大大地减小基于特征分解的信号处理方法的计算量。大量的计算机仿真和水池数据处理结果验证了 SMSWF算法的优越性能。2.在工程应用中,声矢量阵往往存在阵元姿态误差,为了充分了解和校正阵元姿态误差,理论分析了阵元姿态误差对声矢量阵波束图的影响,并就阵元姿态误差对MUSIC算法的影响做了仿真分析。在实际的工作环境中,由于有源校正算法经常受限,提出了一种声矢量阵阵元姿态误差自校正算法,该算法能够实现阵元姿态误差参数和信源DOA的联合估计,且具有良好的参数估计精度和较快的收敛速度。理论分析及计算机仿真验证了该自校正算法的优越性能。3.针对声矢量阵高分辨方位估计算法运算量大的问题,提出了 V-MSWF算法和PV-MSWF算法。V-MSWF算法是标量阵MSWF算法在声矢量阵中的扩展,PV-MSWF算法是基于声压振速联合信息处理,选择参考阵元的电子旋转矢量作为期望信号,运用多级维纳滤波器(MSWF)对信号子空间进行快速估计,该算法基于矢量传感器声压与振速的相干性原理,充分利用了声压振速组合抗干扰能力,有效地抑制了各向同性噪声。这两种算法都不需要计算声矢量阵的互协方差矩阵,不用进行特征值分解,从而大大减小了计算量。计算机仿真及水池试验结果验证了 V-MSWF算法和PV-MS WF算法的DOA估计性能。4.基于声矢量阵的PVFS(Particle Velocity Field Smoothing)算法是一种有效的相干源DOA估计算法,但当存在大量相干源时,该算法性能急剧下降甚至失效。在PVFS算法的基础上,提出了 MSS(Matrix Square Smoothing)-PVFS算法,该算法是对PVFS算法的改进,通过对PVFS算法构造的数据协方差矩阵进行平方、矩阵分块以及矩阵相加等数学运算,增强了 PVFS算法解相干的能力,并大大增加了其分辨相干源的数目。计算机仿真结果及水池试验结果表明,MSS-PVFS算法具有良好的相干源DOA估计性能。
[Abstract]:Compared with the acoustic pressure array, the acoustic vector array overcomes the fuzzy problem of the left and right sides, extends the aperture of the base array, reduces the interference background and the detection threshold, and has obvious advantages in the case of the on-line array, the plane array or the small array element number. Based on the above points, the acoustic vector array signal processing technology is one of the hot spots in the underwater acoustic field, and the United States and Russia strongly study how to apply the acoustic vector array to naval equipment for the detection and localization of underwater targets. This paper mainly studies the correction of the acoustic vector array and the direction-finding technique. The high resolution performance of a large number of acoustic vector array direction finding techniques is obtained on the premise of the ideal array flow pattern, and in the actual situation, the acoustic vector array often has a plurality of array errors, and the array errors will cause the high-resolution direction-finding algorithm performance of the acoustic vector array to be severely degraded and even become invalid. Therefore, it is essential to perform array correction on the acoustic vector array before using the high-resolution direction-finding technique of the acoustic vector array. In the case of acoustic vector array direction finding, most of the current position estimation algorithms of the acoustic vector array have a large amount of computation, which greatly increases the cost and burden of the system and becomes the bottleneck of some systems engineering. In addition, in the underwater acoustic environment, there is often a coherent source, and it is of great significance to study the multi-object resolution of the coherent source of the acoustic vector array to the acoustic vector array direction finding technology. The main contents of this paper are as follows: 1. Based on the multi-level Wiener filter (MSWF), an SMSWF (Simplified Multi-Stage Wiener Filter) algorithm based on multi-level Wiener filter (MSWF) is proposed. The SMSWF algorithm estimates the array error parameters by using the orientation of the correction source and the waveform information, and does not need the covariance matrix calculation and the characteristic value decomposition process, greatly reduces the calculation amount, and has the same array error parameter estimation performance as the characteristic decomposition method. It is found that when a single source is incident on the array and the source waveform is known, the signal subspace obtained by the SMSWF algorithm is equivalent to the signal subspace estimated by the feature decomposition method, so that the calculation amount of the signal processing method based on the feature decomposition can be greatly reduced. A large number of computer simulation and pool data processing results verify the superiority of the SMSWF algorithm. In the engineering application, the acoustic vector array often has a matrix element attitude error. In order to fully understand and correct the matrix element attitude error, the influence of the array element attitude error on the beam pattern of the acoustic vector array is analyzed, and the influence of the matrix element attitude error on the MUSIC algorithm is simulated and analyzed. In the actual working environment, since the active correction algorithm is often limited, a self-correction algorithm for the attitude error of the acoustic vector array is proposed, which can realize the joint estimation of the matrix element attitude error parameter and the source DOA, and has good parameter estimation precision and fast convergence speed. The superiority of this self-tuning algorithm is verified by theoretical analysis and computer simulation. In this paper, a V-MSWF algorithm and a PV-MSWF algorithm are proposed to solve the problem of the high-resolution position estimation algorithm of the acoustic vector array. The V-MSWF algorithm is an extension of the scalar array MSWF algorithm in the acoustic vector array. The PV-MSWF algorithm is based on the combined information processing based on the sound pressure vibration speed, the electronic rotation vector of the reference array element is selected as the desired signal, and the signal subspace is quickly estimated by using the multi-level Wiener filter (MSWF). The algorithm is based on the principle of the coherence of the sound pressure and the vibration velocity of the vector sensor, and makes full use of the combined anti-interference ability of the sound pressure vibration speed, and effectively suppresses the isotropic noise. the two algorithms do not need to calculate the cross-covariance matrix of the acoustic vector array, and the characteristic value decomposition is not needed, so that the calculation amount is greatly reduced. The results of computer simulation and pool test verify the DOA estimation performance of the V-MSWF algorithm and the PV-MS WF algorithm. The PVFS (Particle Velocity Field Smoothing) algorithm based on the acoustic vector array is an effective coherent source DOA estimation algorithm, but when a large number of coherent sources are present, the performance of the algorithm is drastically reduced or even ineffective. On the basis of the PVFS algorithm, a matrix-square-Smoothing-PVFS algorithm is proposed, which is an improvement of the PVFS algorithm. and greatly increases the number of its resolved coherent sources. The results of computer simulation and pool test show that the MSS-PVFS algorithm has good DOA estimation performance.
【学位授予单位】：哈尔滨工程大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN911.7;TN713

【相似文献】