节点文献

双基地雷达弹道目标成像与特征提取方法研究

Research on Bistatic Radar Imaging and Feature Extraction of Ballistic Targets

【作者】 艾小锋

【导师】 肖顺平;

【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】 雷达作为弹道导弹防御系统的核心传感器,在弹道导弹防御的预警探测、跟踪、成像、特征提取与识别、制导、拦截以及毁伤评估等各个环节均发挥着不可替代的作用。攻防对抗的竞争日益激烈,突防措施的不断更新,使单基地雷达面临前所未有的挑战。双/多基地雷达因具备反隐身、抗干扰、信息融合等天然优势,而重新受到重视,特别是随着数字通信、高速运算、频率综合技术的进步,双/多基地雷达的优势逐渐得以实现。本文以双/多基地雷达在弹道导弹防御中的应用需求为背景,深入研究了双基地雷达弹道目标成像与特征提取技术,重点开展了弹道目标双基地散射特性分析与动态回波建模,双基地一维成像与特征提取,二维成像与特征提取三方面的工作:1.弹道目标双基地散射特性分析与动态回波建模采用等效电磁流法求解了高频条件下旋转对称弹头目标的双基地散射特性,证实旋转对称目标双基地散射中心与双基地角平分线上单基地雷达观测的散射中心位置是一致的,采用快速多层多极子电磁计算方法获得了几类弹头模型的宽带双基地散射幅度,验证了上述结论;进一步研究表明旋转对称目标的双基地散射中心随着目标运动、发射站和接收站位置的变化而在目标上滑动;基于滑动型双基地散射中心模型,建立了各个散射中心随目标平动、自旋以及进动的运动模型,为弹道目标散射中心运动规律提供了准确的数学描述;最后结合弹道目标运动模型和电磁计算数据,给出了弹道目标宽带双基地雷达动态回波建模与仿真方法,为弹道目标成像与特征提取提供了数据支撑。2.基于双基地一维距离像的特征提取建立了高速运动目标宽带双基地雷达回波模型,其形式与单基地雷达回波相同,但物理内涵不同,分析了速度、加速度以及双基地角对双基地一维距离像的影响,采用基于多分量多项式相位信号参数估计方法实现了运动补偿;利用补偿后的一维距离像序列构建了时间-距离分布图,提出了基于时间-距离分布图相关的进动频率和剩余速度估计方法,能获得稳定的估计结果;然后结合弹道目标的对称特性,采用扩展Hough变换方法同时获得了完整的锥形目标结构与微动特征,一定条件下解决了弹道目标微动与结构参数耦合难题。利用宽带T/R-R双基地雷达同时观测到目标的单/双基地一维距离像中散射中心的位置,结合目标定位方法,提出了目标长度和姿态角的估计方法,并分析了目标位置、基线长度对估计精度的影响,有效解决了单基地观测一维距离像长度敏感于目标姿态的问题;建立了多基地进动目标观测模型,提出了基于双基地观测的弹道目标进动角和真实长度的联合提取新方法,并讨论了布站对特征提取精度的影响,有效解决了单基地雷达观测信息不完全目标真实长度难以在短时间内获取的问题;提出了基于T/R-R双基地雷达同时观测自旋目标的特征提取方法,两个接收站通过对一维距离像序列的扩展Hough变换处理分别估计参数,然后进行参数关联和融合处理,获得了目标的真实尺寸,解决了单基地雷达观测的旋转半径通常小于真实值的问题。3.双基地ISAR成像与特征提取针对窄带单基地雷达自旋目标二维成像结果受旋转矢量与雷达视线夹角的调制而小于实际尺寸的问题,提出了基于窄带T/R-R双基地雷达自旋目标联合成像方法,通过联合参数估计将成像结果投影到旋转平面上,获得了反映目标真实尺寸的单/双基地二维像,有利于目标识别。建立了基于目标进动的宽带双基地ISAR成像模型,推导了最佳成像时间,通过T/R-R双基地雷达联合观测先估计进动参数,然后选择最佳时间窗内的一维距离像,采用FFT处理可获得清晰的双基地ISAR像;在进动参数未知时,提出先采用时频分析方法获得距离-瞬时多普勒(RID)像序列,利用两个接收站同时观测的单/双基地RID像序列,通过图像配准实现了单/双基地ISAR像的同时横向定标,解决了单基地雷达通过图像配准受进动角大小限制的问题,最后提出了利用定标前后的ISAR序列中特显点的运动规律估计进动角的方法。

【Abstract】 Radar is the key sensor of the ballistic missile defense (BMD) system and it playsan important role in detection, tracking, imaging, feature extraction and recognition,guiding, interception and damage assessment of ballistic missiles. The rivalry betweenballistic missile attack and defense becomes more and more drastic. The means ofjamming and decoying become more and more complex, which makes the monostaticradar meet a great challenge. Bi-and multi-static radar systems have the ability to detectstealth target, anti-jamming and information fusion, so they regain more attention.What’s more, new developments in frequency synthesis, digital communications, andhigh-speed computing allow us to utilize these advantages. With the desirability ofapplying the bi-and multi-static radar systems for BMD, This paper studies the bistaticimaging and feature extraction method of ballistic targets. After analyzing the bistaticscattering characteristics of ballistic targets, the dynamic echo modeling in widebandbistatic radar, bistatic one-dimensional profiling and feature extraction, two-dimensionalimaging and feature extraction methods are investigated. The main contents of thisdissertation include:1. Bistatic Scattering Characteristics Analysis and Dynamic Echo Modeling ofBallistic Targets.With the equivalent electromagnetic current method, the bistatic scatteringcoefficients of rotationally symmetrical warhead model are derived in the highfrequency region and it proves that the bistatic scattering centers of this type target arelocated at the same places as the monostatic scattering centers observed on the bisectorof the bistatic angle. The wideband bistatic scattering coefficients of several types ofwarhead models are obtained via the multilevel fast multipole method electromagneticmethod and the above conclusion is validated. The bistatic scattering centers slip on thetarget when the target, transmitter and receiver is moving, so they are named as slipperybistatic scattering centers. Based on the type of scattering center model, the movingmodel of each scattering center is derived with the target translation, rotation andprecession, which provides an exact mathematic description of the moving rule of eachscattering center. Then, the dynamic echo modeling and simulation method of ballistictargets in wideband bistatic radar is introduced based on the moving model andelectromagnetic calculation data.2. Feature Extraction Based on Bistatic High-Range Resolution Profiles(HRRPs)The echo of the high speed target in wideband bistatic radar is modeled and it is thesame as the monostatic case formally, while the physical meanings are different. Theeffect on the bistatic HRRP of the velocity, acceleration and bistatic angle is analyzed. Then motion compensation is performed based on the multi-component polynomialphase signal frequency estimation and the time-range distribution image is constructedwith the HRRP sequence. A method of estimating the precession frequency and residualvelocity is proposed based on mutual correlation of the time-range distribution and thestable estimation results can be achieved. Afterwards, the whole structure andprecession features are simultaneously extracted by utilizing the symmetricalcharacteristics of ballistic targets and extended Hough transform (EHT), which providesa solution to the coupling problem of micro-motion and structure parameters undersome conditions.In the wideband T/R-R radar system, the T/R and R receivers obtain the lengthsand the locations of the monostatic and bistatic HRRPs separately. Combined with thetarget localization method, the target length and attitude angle are estimated, and thesimulation results are provided for analyzing the effect of the target location andbaseline length on the estimation errors. This is helpful for overcoming the problem thatthe length of the monostatic HRRP is sensitive to the target attitude. Then, the model ofmulti-aspect observation is introduced and a novel algorithm, which is based on themulti-aspect observed HRRP sequences, is proposed to estimate the precession angleand real target length jointly. And the effect of the radars’ distribution on theestimated errors is discussed. For the spinning targets, a method for feature extraction isproposed based on T/R-R bistatic radars. In this method, each receiver estimates someparameters via EHT firstly, then parameter association and fusion processing areperformed, finally the real rotational radius is obtained. This method is helpful forovercoming the problem that the extracted rotational radius via monostatic radar isusually shorter than the real value.3. Bistatic Two-dimensional (2-D) ISAR Imaging and Feature ExtractionThe2-D images of spinning targets obtained via the narrowband monostatic radarare always smaller than the real size. So, a novel2-D imaging algorithm is presentedbased on the connections between the mono-and bistatic echoes of the same scattererand the classical Hough transform (HT), allowing both the mono-and bistatic2-Dimages projected on the spinning plane to be obtained simultaneously. Both mono-andbistatic2-D images reflect the real size of the spinning target, so it is helpful targetrecognition.The bistatic ISAR imaging model of precession target is introduced and theoptimum imaging time is derived. Afterwards, through selecting the optimum bistaticHRRPs, a clear bistatic ISAR is obtained via the classical fast Fourier transformprocessing on condition that the procession parameter is estimated based on the mono-and bistatic HRRP sequences and EHT. When the precession parameters are unknown,the time-frequency transform method is proposed for producing the mono-and bistatic range instantaneous Doppler (RID) image sequences, then both mono-and bistaticISAR images are scaled via image registration of mono-and bistatic RID images. Thismethod overcomes the problem that the image registration precision is limited by theprecession angle in the monostatic radar. Finally, a novel method is proposed forestimating the procession angle by utilizing the moving rule of the dominant scatteringcenter in the un-scaled and scaled ISAR image sequence.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络