机载认知雷达中的KA-STAP研究

发布时间：2018-01-17 23:01

本文关键词：机载认知雷达中的KA-STAP研究　出处：《电子科技大学》2017年博士论文　论文类型：学位论文

【摘要】：认知雷达是一种知识辅助的闭环雷达体制,具有根据环境感知信息、先验知识以及任务特点对雷达发射信号以及接收端信号处理方式进行优化的能力。将认知雷达的工作模式引入机载雷达系统,可提升空时自适应处理算法在非均匀杂波背景下的目标检测性能。知识辅助空时自适应处理(KA-STAP)是机载认知雷达信号处理中的关键环节。本文围绕机载认知雷达中的KA-STAP技术展开研究,主要工作和贡献为:(1)研究了机载雷达的杂波建模方案环境杂波建模是机载雷达信号处理算法研究的重要辅助工作之一。本文给出了基于积分式杂波模型的机载雷达非均匀杂波建模方案。针对已有海杂波混沌建模方案难以稳定拟合实测海杂波序列的问题,对实测海杂波序列与混沌序列进行了对比研究,提出了一种海杂波信号分部式分解模型(SCD)。在该分解模型的基础上给出了海杂波的分部式拟合方案(SCCS)。与已有的海杂波混沌拟合方案相比,SCCS拟合方案对不同海态条件下的海杂波序列具有稳定的拟合能力,为海杂波混沌建模问题提供了新的研究角度。(2)研究了锥化矩阵模型及其改进方案机载认知雷达中的知识辅助空时自适应处理(KA-STAP)的工作原理是利用先验的杂波协方差矩阵对样本估计协方差矩阵进行修正,以增加空时自适应滤波器在非均匀杂波环境下的输出性能。锥化矩阵是先验杂波协方差矩阵中的一个子成分,作用是调整先验杂波信号在角度-多普勒域的二维谱结构,其取值将直接影响KA-STAP滤波器的凹口形状。常用的CMT锥化矩阵模型仅包含一个待定参数,可优化自由度较低,对海杂波谱的拟合能力有限。本文在CMT模型的基础上,提出一种推广的锥化矩阵模型(GCMT)。针对海杂波谱的结构特征构造了用于中高频段谱宽调整的带通分布随机相位导向矢量,并将待优化参数由原CMT模型中的辛格函数主瓣宽度参数变为多个子成分的加权系数。与CMT模型相比,GCMT模型增加了锥化矩阵模型的可优化自由度,提升了先验协方差矩阵的拟合能力。(3)研究了锥化矩阵待定参数的自适应优化算法给出了锥化矩阵待定参数的两种自适应优化算法。算法一借鉴了已有的PW法色加载系数优化算法的优化策略,针对锥化矩阵参数优化问题的特点设计了不同于PW法的预白化性能评估量,构造了以先验协方差矩阵预白化性能最大化为优化目标的锥化矩阵参数优化函数(简称为TPO算法)。TPO算法同时适用于空域维与时域维锥化矩阵的优化。算法二将接收阵列在慢时间维与快时间维的采样数据同时用于杂波方位维最小方差谱的估计,并以先验数据与实测数据的谱相似度最大化为目标设计优化函数。两种优化方案均将实时回波信号引入了先验信息的优化,增加了知识辅助杂波抑制算法对时变杂波环境的适应能力。(4)色加载系数优化算法的稳健性设计色加载系数是KA-STAP滤波器权矢量计算式中的待定加权系数,其取值决定了先验信息在滤波器权矢量设计中所占的比重。色加载系数的大小应由先验信息的准确度以及样本估计矩阵对真实杂波加噪声协方差矩阵的逼近程度共同决定。已有的基于预白化性能评估的色加载系数优化算法(PW法)无法对当前待检测单元(CUT)的先验信息进行准确度评估,并且无法对每一个待检测单元得到独立的色加载系数优化结果,因此对于先验信息在距离维的非均匀性不具有鲁棒性。本文针对PW法存在的鲁棒性问题给出一种改进方案(简称CPW算法)。CPW算法能够根据当前待检测单元先验信息的准确度调整色加载系数的取值,并且能够对每一个待检测单元得到独立的色加载系数优化结果。在与降维KA-STAP结合的情况下,CPW算法对于先验信息在距离维以及多普勒维的非均匀性均具有鲁棒性。(5)研究了低秩矩阵元素补齐技术在KA-STAP中的应用由于CUT单元的回波可能包含目标信号,已有的KA-STAP算法均没有将其用于滤波器的设计。为了提取CUT单元的杂波信息,本文提出一种CUT单元目标信号剔除算法。该算法利用了机载雷达空时二维杂波谱矩阵的低秩性,以及点目标信号在角度-多普勒域内仅分布于少数通道的特点,将低秩矩阵元素补齐技术用于CUT单元杂波信号的恢复,设计了一种在变换域进行的CUT单元目标信号剔除算法。以锥化矩阵参数优化问题为例,给出了该算法在KA-STAP中的应用方式。CUT单元杂波信号的引入降低了KA-STAP中各参数优化问题对参考单元样本的依赖性,增加了KA-STAP对样本筛选算法性能不佳问题的鲁棒性。(6)结合实测数据的KA-STAP有效性测试利用机载相控阵雷达实测杂波数据对KA-STAP算法的有效性进行了测试。设计了脱离地形数据库的全自适应KA-STAP处理流程,包括环境信息提取、先验杂波协方差矩阵参数优化,以及色加载系数优化等环节。仿真测试验证了本文所述部分理论和算法的有效性。测试结果表明在样本个数低于两倍系统自由度的条件下,KA-STAP滤波器具有优于常规STAP滤波器的输出性能。
[Abstract]:Cognitive radar is a kind of knowledge aided closed-loop radar system is according to the environmental perception of information, knowledge and task characteristics prior to transmitting and processing method of radar receiver optimization capability. The cognitive radar working mode in airborne radar system, which can enhance the space-time adaptive processing algorithm in inhomogeneous clutter background the target detection performance. Knowledge aided space-time adaptive processing (KA-STAP) is a key step in the signal processing of airborne cognitive radar. This paper focuses on the research of airborne radar in cognitive KA-STAP technology, the main work and contributions are: (1) study of the airborne radar clutter modeling scheme of clutter modeling is an important airborne radar signal processing the auxiliary algorithm presented in this paper. One of the integral type airborne radar clutter wave model of inhomogeneous clutter modeling scheme based on the sea clutter. Chaos modeling scheme is difficult to stabilize the fitting of the measured sea clutter sequence, the measured sea clutter sequence and chaotic sequence are studied, put forward a kind of sea clutter signal segment decomposition model (SCD). Based on the decomposition model is given on the sea clutter segment fitting scheme (SCCS) compared with the existing sea clutter chaos fitting scheme, SCCS fitting scheme for different sea state conditions of the sea clutter sequence has stable fitting ability, provide a new angle for research of sea clutter chaos modeling problem. (2) studied the cone matrix model and the improvement scheme of airborne cognitive knowledge aided space in radar time adaptive processing (KA-STAP) the working principle is to use a priori the clutter covariance matrix of sample covariance matrix is modified to increase the space-time adaptive filter in inhomogeneous clutter environment. The output performance of cone The matrix is a priori a sub component of the wave clutter covariance matrix, is to adjust the prior complex two-dimensional wave signals in the angle Doppler domain spectrum structure, its value will directly affect the KA-STAP filter notch shape. CMT taper matrix models contain only one undetermined parameter, can optimize the degree of freedom is low, the fitting the ability of sea clutter. Based on the CMT model, we propose a generalized cone matrix model (GCMT). According to the characteristics of sea clutter spectrum is constructed for bandpass random phase distribution of the steering vector of high frequency Pukuan adjustment, and the parameters to be optimized by the Singer function of the main lobe width parameter the original CMT model into several sub weighting coefficient components. Compared with the CMT model, the GCMT model can increase the degree of freedom optimization of taper matrix model, to enhance the capability of fitting a priori covariance matrix. (3) studied the taper The matrix parameters of the adaptive optimization algorithm given taper matrix parameters of two kinds of adaptive optimization algorithm. The algorithm uses a known PW method of color optimization algorithm of loading coefficient, designed for cone whitening properties different from PW method evaluation of characteristic matrix parameter optimization problem, constructed with the largest the prior covariance matrix pre whitening performance into function optimization of matrix parameters of cone optimization target (called TPO algorithm).TPO optimization algorithm is also applicable to the spatial dimension and time dimension taper matrix. The algorithm two will receive array in the sampling data in slow time dimension and time dimension at the same time for fast estimation of clutter azimuth minimum variance spectrum, and spectral similarity to maximize the prior data and the measured data for the optimization of objective function design. Two optimization schemes are introduced to optimize real-time echo signal prior information, Increase the knowledge aided clutter suppression algorithm for time-varying clutter environment. (4) robust design color color loading coefficient loading coefficient optimization algorithm is to calculate the weighted coefficient in the pending KA-STAP filter weight vector, its value determines the percentage of prior information in the right vector in the design of filter size by proportion. A priori information color loading coefficient accuracy and sample estimation matrix of real clutter plus noise covariance matrix approximation of the joint decision. Color loading coefficient evaluation and optimization algorithms based on the pre whitening performance (PW) method on the current detecting unit (CUT) of prior information on the accuracy assessment, and not to each test unit can load coefficient optimization results independently, so the prior information in the range of the non uniform is not robust. Aiming at the defects of PW method in Shandong The robustness is given an improved method (CPW algorithm).CPW algorithm can adjust the color value according to the accuracy of the loading coefficient detection unit for the prior information, and can get the optimal color loading coefficient results of independent for each unit to be detected. In combination with reduced dimension KA-STAP, CPW algorithm for a priori information in the range of uniform dimension and Doppler are non robust. (5) on the low rank matrix elements to complete the technology in the application of KA-STAP as the CUT unit may contain the target echo signal, the KA-STAP algorithm has no be used for filter design. In order to extract the clutter information of the CUT unit in this paper, a CUT unit target signal eliminating algorithm. The algorithm uses the low rank of the airborne radar space-time clutter spectrum matrix, and point target signal in angle Doppler domain only In a small number of channels of cloth will low rank matrix elements filled technique used in CUT clutter signal recovery unit, a CUT unit in the transform domain signal elimination algorithm design. Optimization problems for example to taper matrix parameters, gives the.CUT application of the algorithm in the KA-STAP unit in the clutter was introduced the signal is reduced by KA-STAP in the parameter optimization problem of reference unit sample dependence, increase the robustness of the algorithm performance of KA-STAP screening samples. (6) the effectiveness of combining effectiveness test using ray airborne phased array KA-STAP data of the measured clutter data of the KA-STAP algorithm is tested is designed. From the topographic database process fully adaptive KA-STAP, including environmental information extraction, a priori clutter covariance matrix of parameter optimization, and color loading coefficient optimization aspects simulation testing. The effectiveness of the theory and algorithm described in this paper is tested. The test results show that the KA-STAP filter has better output performance than the conventional STAP filter under the condition that the number of samples is less than two times the degree of freedom of the system.

【学位授予单位】：电子科技大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TN957.51

【相似文献】