【作者】 刘俊凯；

【导师】 王雪松；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2012， 博士

【摘要】 飞机尾流是飞机产生升力的必然产物。与周围大气相比，尾流具有独特的空气动力学特性和雷达特性。从空气动力学特性看，尾流内空气流动非常强烈而且不规则，是航空安全的重要影响因素。从雷达探测的角度看，尾流具有较强的雷达可观测性以及独特的多普勒和空间分布特性。尾流雷达探测是航空安全和反隐身等领域的前沿研究课题，论文以此为背景，研究尾流雷达回波模拟、尾流雷达探测和尾流雷达跟踪等问题，主要成果包括以下三个方面：首先，对潮湿大气条件下飞机尾流的雷达回波进行建模仿真分析，基于尾涡速度模型，提出了雷达视频回波建模方法和统计直方图方法，得到了相关处理间隔内的飞机尾流的脉冲回波序列和尾流谱解析表达式，研究结果表明，尾流回波具有独特的频域扩展特性，与雷达探测实验结果相吻合，为飞机尾流探测提供了回波模型基础。其次，针对具体的飞机尾流探测场景和应用需求，提出三种尾流检测方法。面向长驻留时间情况下尾流监视需要，提出基于多普勒特性的扩展目标CA-CFAR检测方法，给出了高斯分布和Gamma分布杂波背景下虚警概率和检测概率的解析表达式，仿真实验表明，通过自适应选择最优检测单元个数，该方法的检测性能优于传统尾流检测方法；面向短驻留时间条件下尾流监视需要，提出了一种基于短脉冲回波序列的信息几何检测方法，仿真分析了高斯杂波环境下的检测性能，仿真实验表明，信息几何方法的检测性能优于常规相参积累方法；面向复杂天气条件下尾流监视需要，提出了一种基于神经网络的尾流检测方法，通过利用尾流多普勒谱特性来训练后向传播神经网络，可有效实现气象杂波背景下的尾流检测。第三，针对尾流回波独特的空间线状分布特性，面向机械扫描脉冲多普勒雷达大空域飞机尾流监视需求，提出了飞机尾流检测前跟踪方法和跟踪方法。在检测前跟踪阶段，该方法通过对特定空间分辨单元进行频域低门限检测，然后针对尾流检测点迹的空间分布特点建立系统状态方程，提出了结合幅度与位置量测的检测前跟踪新方法。该方法可容许预处理较高的虚警，通过空间积累实现低虚警。在跟踪阶段，提出基于空间线状分布模型的尾流跟踪方法，能估计出尾流的中心、倾角、长度等状态向量参数，可实现尾流连续跟踪。论文的研究成果对提高军用/民用飞机的飞行安全、增强军用飞机在应急条件下的快速出动能力、提升军用雷达探测隐身目标的能力具有重要的理论意义和军事、民事应用价值。更多还原

【Abstract】 The wake vortices are the inevitable phenomena of aircraft lift. Compared with theambient atmosphere, the wake vortices have particular aerodynamic characteristics andradar characteristics. For the aerodynamics, the eddying motion of the atmosphere in thewake is very strong and irregular. It is one of the main factors related to aviation safety.For the radar detection, the wake vortices have relatively stronger detectability along withunique Doppler and spatial distribution characteristics. The wake vortices detection is anadvanced field of aviation safety and anti-concealed. Facing the need of aviation safety andanti-concealed, radar echo modeling of the vortices, radar detection, and tracking areinvestigated in the thesis. The main research results are listed as the following threeaspects:First, radar echo modeling of vortices in the condition of wet atmosphere is analyzed.Two kinds of the signal modeling methods are put forward. The pulse sequence in a CPIcan be obtained using the radar video frequency echo model, and the analytical expressionof the spectrum can be acquired using the statistics histogram method. The research resultsshow that the vortex echoes have unique Doppler-spread characteristics. It is identical tothe experiment results.Second, three vortices detectors are put forward based on the particular explorationand application situation. For the need of wake vortices surveillance in long dwell timecondition, a frequency-domain distributed target detector is put forward, and the analyticalexpressions of false alarm probability and detection probability in Gauss and Gammadistribution clutter are obtained. Simulation experiments show that through adaptivelyselecting the optimal number of detection cells, the detection performance of the proposedmethod is better than the conventional vortices detection method. For the need of wakevortices surveillance in short dwell time condition, an information geometry detector basedon short pulse sequence is put forward, and the detection performance in Gauss clutter wasanalyzed. Simulation experiments show that the detection performance of the informationgeometry method is better than the conventional coherent accumulation method. For theneed of wake vortices surveillance in the complex weather condition, a wake vorticesdetection method based on neural network is proposed. Through training theback-propagation network using the vortex spectrum characteristics, the method can detectthe wake vortices effectively in the complex weather clutter.Third, for the need of wake vortices surveillance using the mechanical scan Dopplerradar in the large airspace, based on the spatial linear distribution characteristics, a TBDmethod and a tracking method of airplane wake vortices are proposed. In TBD phase, firstof all, the frequency-domain low-threshold detection is carried out, so the range-angle-Doppler measurement space is reduced to two dimensions of range-anglespace. Then the system state function is set up based the spatial distribution characteristicsof detection points, and a new particle filter TBD method which combines amplitude andposition measurements is proposed. The TBD method tolerates high level of false alarms inpreprocessing, and achieves low false alarms through spatial accumulation. In trackingphase, a wake vortices tracking algorithm based on the spatial linear distribution model isput forward. Wake state parameters, such as center, slope angle, and length can beestimated by this algorithm. The influence of the wake measurement number variety iseliminated using the tracking method, and continuous tracking of the wake vortices isrealized also.Research results of this thesis are of important theory meaning and application valueto improve flight safety of fighter plane and civil aircraft, enhance fast takeoff of fighterplane in an emergency, and improve the detection performance of concealed targets.更多还原