【作者】 贺思三；

【导师】 付强；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2010， 博士

【摘要】 一般把目标或目标组成部分除主体平动之外的振动、转动等小幅运动称为微动。微动在自然界广泛存在并反映了目标一些特有的运动及结构特征,在雷达目标探测与识别中受到广泛关注。本文针对高分辨体制下的微动目标回波信号处理进行研究,开展的工作和取得的成果主要包括如下几方面:第二章系统研究了微动目标的运动模型及其对高分辨雷达的调制效应。首先建立了目标的三维运动模型,然后结合线性调频(LFM)及频率步进(SF)两种典型高分辨体制分析了微动对多普勒频率、距离像、距离像序列及二维像的影响,并针对SF雷达体制距离-多普勒耦合严重的问题,提出了等效瞬时多普勒频率的概念。第三章对微多普勒信号与目标主体回波信号分离展开研究。首先阐述了基于调频率参数微多普勒信号分离的基本原理;然后对基于Chirplet分解的微多普勒信号分离算法展开研究,提出了一种快速Chirplet参数估计算法,并推导了其克拉美罗限(CRB),分析了各参数变化对调频率参数估计性能的影响;为进一步减少计算量,提出了一种基于调幅-线性调频(AM-LFM)分解的微多普勒信号分离算法,该算法对信号包络形式没有严格限制,具有运算量少、分离效果好的优点。第四章对典型微动结构的参数估计展开研究。通过对高分辨体制下(LFM、SF)振动及旋转微动结构回波模型的分析,指出它们的调制效应表现为二维图像中(时间-距离像、时间-频率像)的正弦曲线,推导了正弦曲线参数与微动参数之间的对应关系,提出了一种基于分步参数估计的正弦曲线快速提取算法。对中段弹头这一特殊微动目标,就LFM体制下的进动参数估计展开研究,提出了基于成像质量的进动参数估计算法,并设计了搜索参数降维流程:通过对回波信号对称轴位置的估计将搜索参数维数降至二维,并进一步利用弹头目标的结构特性及散射特性将搜索参数降至一维。该算法合理挖掘了弹头的结构特性及散射特性,可在未知目标结构尺寸条件下在一个进动周期内实现进动参数的估计。第五章对小角度微转动目标的高维成像进行研究。目标转动产生的多普勒信息是二维成像的基础,但传统成像算法只能对平面内匀速转动目标成像。通过对海面颠簸舰船回波相位的分析,本章提出利用如下调幅-调频模型对复杂运动目标的雷达回波进行建模:利用Q阶多项式对调幅函数进行描述,利用包含高阶误差项的三次多项式对相位进行描述。针对上述模型,提出了参数估计算法,得到了目标上多个强等效散射中心的相位历程,组成相位矩阵。然后通过对相位矩阵的分析,提出了基于相位矩阵奇异值分解的小角度微转动目标高维成像方法:对平面非匀速转动目标可得到聚焦的二维像,对三维空间转动目标可得到失真的三维散射中心模型。第六章总结了论文的研究工作和主要创新点,指出需要进一步研究的问题。更多还原

【Abstract】 Generally, micro-motion is defined as the mechanical vibrations or rotations or other high-order motions of a target or any structure of target in addition to its bulk translation. Micro-motion is widespread in nature and reflects some unique movement and structure character of objects. Therefore, it draws great attentions to the researchers from the areas of target detection and recognition. This dissertation researches the problem of how to deal with the high resolution radar echos from micro-motion targets, and the main studies and contributions can be summarized as follows:In chapterⅡ, the motion model for micro-motion targets and its modulation effect for high-resolution radar is studied systemly. First of all, the three-dimensional motion model of target is established. Then, combined with the two typical wideband waveform, LFM signal and SF signal, the modulation effects caused by micro-motion on Doppler frequency, high resolution range profile, profile sequence and ISAR image are analyzed. Especially, the equivalent instantaneous Doppler frequency is introduced to deal with the coupling between range and Doppler for SF wideband radar.The issue of micro-Doppler signal separation for targets with micro-motion structures is addressed in chapter III. Firstly, the basic principle of signal separation based on chirp rate is introduced. Then, a fast algorithm on parameter estimation of chirplet signlas is proposed. Also, the CRB on Chirplet parameter estimates is deduced and the influence of other parameters on the Chirp rate estimate is analysed. To further reduce computation, a micro-Doppler signal separation method based on AM-LFM decomposition is proposed, which has no constraint on signal envelope and can obtain a better separation results with less computation.The parameter estimation for some typical micro-motion structure is studied in chapter IV. It’s pointed out that the vibration and rotation structures will appear as a sinusoid in two-dimensional image (time-range image or time-Doppler image) under the LFM or SF wideband systems. The relationship between the sinusoid parameters and micro-motion parameters is deduced and a time saving approach to extract the sinusoid is proposed. Then, a method based on imaging quality is proposed to estimate the precession parameters for the precessing warhead in middle course. Also, an approach is designed to reduce the searching parameters for the proposed method: searching parameters is reduced to two dimensions after estimating the symmetry axis of echos, and they are further reduced to one dimension by ultilizing the structural and scatter characters of warheads. This algorithm exploits the warheads’s structural and scatter characters reasonably, which can estimate the precssion parameters of warheads with unknown size during one period.ChapterⅤstudies how to obtain high-dimensional image for small-angle micro-rotation targets. The Doppler frequency casued by rotation is the foundation for two dimensional imaging, but traditional imaging algorithms only suit for the uniform rotation situation. Taking the oscillating ships in sea as an example, this chapter proposes to use the following AM–FM function to model the radar echoes of complex moving targets: the amplitude of scattering centers are modeled by Q-order polynomial functions, the phase of scattering centers are modeled by cubic polynomial with higher-order error items. Also, a parameter estimation algorithm is proposed for the above model and the phase matrix is formed after the phase history being established for each scattering center. Then, a method based on the singular value decomposition of the phase matrix is proposed to obtain the image for small-angle micro-rotation targets. The image is a focused two-dimensional image for nonuniform rotating targets in plane or a three-dimensional scattering center model for three dimensional rotation targets.In chapterⅥ, the contributions of this dissertation are summarized and the potential problems and difficulties are also pointed out.更多还原