【作者】 江磊；

【导师】 蔡平；

【作者基本信息】 哈尔滨工程大学 ， 通信与信息系统， 2008， 博士

【摘要】 小尺度阵易于安装,便于使用,越来越受到水下各种设备的青睐。近年来水下系统、设备的小型化趋势,使得人们更多的关注如何充分挖掘、发展小尺度阵的信号处理方法与能力。通常较大的阵列尺度是阵列实现高分辨率的关键因素,而水下平台往往体积有限,使得水下设备往往不能有大的基阵尺度,因此“高分辨”与“小尺度”成为水声设备发展的瓶颈。同时,水下干扰源复杂,往往需要在多干扰的低信噪比环境中对目标进行探测和跟踪。本文正是希望能够探索并发展适用于小尺度基阵的多目标跟踪方法。本文首先从分析均匀线阵的波束域空间功率谱以及相关函数入手,提出一种新的协方差矩阵构造方法,给出波束特征值的概念,并证明波束特征值是由阵列流型所决定的。在此基础上,提出了BEM算法及其具体实现过程。并通过与两种常见的信号源数估计算法:AIC和MDL算法进行比较,发现现有的估计方法适用的信噪比较高,而BEM算法可以在更低的信噪比条件下正确估计信源数。之后,利用虚拟阵元技术与BEM算法相结合,进一步提出适用于小尺度阵的V-BEM信源数判决准则。通过虚拟阵元,改善了波束图的形状,抑制了栅瓣的产生,来获得稳定的波束特征值分布。本文提出的两种算法不仅适用于非相干目标入射的信源数估计,而且适用于相干目标入射的情况,并且避免了由于空间平滑等方法带来的基阵孔径的损失,而基阵孔径的损失对于小尺度阵来说尤其敏感。舰船、潜艇和各种水下潜器,在水下机动时往往都有很强的自噪声,现代声纳系统需要在这样的噪声条件下,对目标进行正确的检测、识别和估计,因此,如何在强干扰的背景下来提高对低信噪比信号的检测能力,成为现代声纳信号处理不得不面对的一个重要课题。同时,小尺度阵平台的多个目标的跟踪能力也倍受关注,期望能够进一步解决强干扰下的弱目标检测问题。因此,本文基于对干扰方向信号进行屏蔽的思想,设计了逆波束形成抵消技术,推导了逆波束形成矩阵和阵列流型之间的关系。给出声压阵相移和时延逆波束形成方法,分析了常规的声压阵仅能够利用声压信息进行抵消,会导致同时消除左、右舷方向的信号。而本文提出的基于矢量阵组合指向性逆波束形成的干扰抵消技术,可以对0°～360°方向内的干扰进行抵消。同时,利用矢量阵的逆波束形成干扰抵消,可以更好的解决:“矢量阵左右舷模糊与多目标辨别”问题。为了使小尺度阵能够达到水雷用声引信及其它水下武器系统对水下多目标同时跟踪的工作需求,设计了基于小尺度矢量阵的多目标跟踪算法。小尺度阵由于受到基阵尺度的限制,具有较宽的旁瓣,而且阵型固定,因此跟踪较高频段的信号时会产生栅瓣。针对这些问题,探讨了在小尺度阵中应用虚元技术的特点。采用了并行多组自适应Notch滤波器划分频带对目标辐射噪声进行跟踪的工作方式,通过虚元和恒定束宽相结合在有效地减小波束主瓣宽度的前提下,克服了栅瓣的影响和相位模糊问题,并且将虚元和自适应Notch滤波器离线重构波束形成、变频恒定束宽以及矢量阵逆波束形成技术相综合,提高了同频段空域目标的估计性能,可以对同频段空域分离的强弱目标进行估计。通过对各种方法的综合利用,使得本文设计的小尺度阵跟踪算法可以有效地对多目标进行跟踪。更多还原

【Abstract】 Small-scale array has the merit to be installed easily, which the underwater vehicles are preferred to equip with. Miniaturizing the system and developing new signal processing methods to promote the ability for the small-scale array are becoming a hotspot in underwater equipment production. The limit of the underwater instruments’ cubage restricts to get large array scale in some application situation, while the large scale is the key to obtain high resolution ability. It is usual to track the target in low signal to noise ratio (SNR) in the presence of many interference sources. This dissertation tries to explore the new methods for the multi-target tracing in small-scale array.The relationship between the beam space power and the correlation function in uniform linear array is analyzed first to propose a new covariance matrix and the idea of the beam space eigenvalue. The distribution of this eigenvalue is determined by the array manifold.Then the beam eigenvalue method(BEM) algorithm is discussed. Comparing with AIC and MDL algorithms, the BEM can estimate the source number in lower SNR. Incorporating with the virtual element methods, the virtual-BEM(V-BEM) algorithm is proposed for the small-scale array source number estimation. The beamforming pattern is improved by adding the virtual element to inhibit the gratinglobe.The distribution of the beam eigenvalue becomes more stable for estimation.The two methods proposed can be used not only for incoherent targets but also for coherent targets, and avoids the aperture loss caused by the space smoothing,which is very important for the small scale array.The self-noise spectrum of the ship, submarines and various underwater vehicles are high. The modern sonar is forced to work and required to promote the low SNR detectation ability in presence of strong interference. At the same time,it is also focused to trace and detect the multi-target with strong interference by using small-scale array. So based on the idea of shielding the interference direction, the inverse beamforming algorithm is designed and the relationship between the inverse beamforming matrix and beamforming matrix is deduced. The inverse beamforming using the phase-shift and time delay is presented. The interference from both larboard and starboard is cancelled simultaneously if just the pressure used; while the vector array inverse beamforming counteracting can cancel the interference in 0°～360°. The problem of distinguishing whether there is multi-target impinging or vague caused by single target could also be solve using the vector array inverse beamforming.To satisfy the work requirement of the acoustic fuze of underwater mine and some other underwater weapons, the multi-target tracking system is designed based on the small-scale vector array. The small-scale array is restricted in array aperture which causes wider side lobes. Using the fixed small array to track higher frequency targets will cause grating lobe. A group of parallel Notch filter with different frequency is used to track ship targets by the radiated noise .Through the combination of the virtual array and constant beam width technique, the gratinglobe and the phase vague problem is overcame. The system uses the virtual elements, Notch filter offline beamforming, constant beamwidth and the vector array inverse beamforming algorithm synthetically to promote the performance of the small scale vector array. The weaker target can be estimated by canceling the strong targets, even if the two targets working in the same frequency. The multitarget tracking performance can be also promoted. The simulation and experiments analysis has confirmed the ability of this system.更多还原