【作者】 江碧涛；

【导师】 姜景山； 张云华；

【作者基本信息】 中国科学院研究生院（空间科学与应用研究中心） ， 计算机应用技术， 2007， 博士

【摘要】 分布式雷达因为在实现高分辨率成像、宽刈幅成像、三维成像和地面慢速运动目标检测等方面具有广阔的前景,因而受到广泛的关注并因此成为雷达技术,尤其是合成孔径雷达(Synthetic Aperture Radar,SAR)技术研究的热点领域。论文针对分布式雷达的高分辨率成像和地面运动目标检测问题开展工作,重点研究分布式SAR的成像算法和分布式稀疏孔径雷达地面运动目标检测(Ground Moving Target Indication, GMTI)STAP算法和扫描方向干涉雷达(Scanned Pattern Interferometric Radar, SPIR)方法。各章节的主要内容如下:第1章介绍本文的研究背景和国内外相关领域的研究现状,归纳了论文的主要工作和贡献。第2章在论述SAR的基本原理的基础上,重点介绍SAR成像模拟方面的工作。研究中观察到当用点目标模拟面目标时,会在SAR幅度图像中出现明暗条纹,在条纹的亮区分辨率和清晰度都很高,相位信息得到很好保持,但在暗区则分辨率和清晰度都明显降低,相位信息也受到破坏。用天线阵的观点对此进行了解释并提出三种消除这一现象的方法。第3章介绍对宽带调频步进信号的两种处理算法,即基于匹配滤波的子孔径方法和基于de-chirp处理的方法,并对这两种方法进行了比较。本章的工作为第4章分布式SAR成像算法研究奠定了理论基础。第4章首先建立分布式SAR距离向和方位向的信号模型与调频步进信号模型的等效关系。但是这种等效关系的建立须满足一定的条件,即在距离向各SAR进行匹配滤波时须以观测区域中的同一参考目标来决定匹配滤波函数参考距离,而在方位向,各SAR的斜视角需较小,且斜视角之间形成一个等差数列关系。本章还提出了广义合成孔径概念,即跨平台进行孔径合成的概念。广义合成的前提是要保证各个平台的SAR所获得的关于同一地面区域的观测频谱是相干的,即经过了相位误差校正。分析表明影响分布式SAR距离向分辨率提高的主要因素是各个SAR系统距离测量的相对精度,而影响方位向分辨率提高的主要因素是斜视角的测量精度。第5章针对分布式稀疏孔径雷达实现GMTI的空时自适应处理(Space Time Adaptive Processing, STAP)算法和SPIR算法进行研究。在STAP算法研究中,重点分析影响STAP算法的主要因素,针对稀疏孔径出现的栅瓣从而导致的速度检测盲区问题,用多载频正交技术和非均匀阵列方法加以有效克服。在SPIR方法研究中,重点讨论了影响SPIR性能的主要因素,即分布式雷达的位置测量精度、收发通道的幅度和相位精度。研究表明SPIR方法尽管可以较好地克服稀疏孔径带来的栅瓣问题,但是受地面杂波的影响较大,而地面杂波的影响从本质上体现在点扩散函数(Point Spread Function, PSF)样本矩阵的条件数增大上,此时必须采用Clean算法来加以克服。模拟研究还表明,随着阵列单元数量的增加,SPIR方法对上述误差的容忍度也随之增加。对于设计好的一定构形的分布式稀疏孔径,在一定误差范围内,准确测量孔径实际位置比精确控制孔径位置更重要。第6章对全文的工作进行回顾和总结,指出论文工作存在的局限并提出在今后应该开展的四个方面的研究工作。更多还原

【Abstract】 Distributed radars have been paid to extensive attention in recent years and are becoming the hot research field in radar technologies, especially in SAR, because of their broad prospects in realizing high-resolution, wide-swath, three-dimensional imaging and GMTI (Ground Moving Target Indication). The dissertation focuses on the problems of imaging and GMTI by distributed radar with emphases on the imaging algorithms of distributed SAR and GMTI algorithms of STAP (Space Time Adaptive Processing) and SPIR (Scanning Pattern Interferometric Radar). The major contents of each chapter are summarized as follows:Chapter 1 introduces the background and current research status in domestic and abroad, summarizes the major works and contributions.Chapter 2 introduces the works on SAR imaging simulation. Study shows that bright and dark strips appear in simulated SAR images, when point-targets are used to simulate distributed target. In the bright strip areas, both of the resolution and definition are very high, and the phase information is well preserved. In dark strips however, both of the resolution and definition are degraded, as well as the phase information is destroyed. Radiation pattern of sparse array antenna is applied to explain the phenomenon and three methods are proposed to remove this effect.Chapter 3 introduces two processing method for stepped-frequency chirp signals (SFCSs), i.e. the subaperture method based on matched filtering and the de-chirp method, and gives comparison between these two methods. The works in this chapter lay a theoretical foundation for the development of imaging algorithm in chapter 4.Chapter 4 firstly establishes an equivalent relation between the signal models of distributed SAR in both range and azimuthal directions and that of SFCSs. Two preconditions for the existence of equivalent relation should be contented, i.e. the same target point in observed area should be selected to calculate the reference ranges for matched filter of each SAR in range direction, and in azimuthal direction, there should be small squint angle for each SAR and the angle difference between any two neighbor SARs is the same. And then, the algorithms introduced in chapter 3 are applied to process the range and azithumal signals of distributed SAR. This chapter also proposes the idea of general aperture synthesis, i.e. the idea of aperture synthesis across different platforms. The presupposition of general aperture synthesis is that the observed spatial spectrum should be coherent with phase differences between different platforms removed. Studies show that the relative accuracy of range measurement between different SARs and the squint angle measurement accuracy of each SAR dominate the facts of preventing resolution improvement in range direction and azimuthal direction, respectively.Chapter 5 concentrates on the STAP and SPIR algorithms for distributed radar with sparse apertures realizing GMTI functions. In the study of STAP, major facts for affecting the performance of STAP are stressed. In view of the grating lobes, i.e. blind zone problems, inherently existed in sparse array, techniques of using multi-carrier frequencies and non-uniform of array are used to mitigate this problem. In the study of SPIR, the key facts influencing the performance of SPIR investigated, such as the position accuracy of array element, the amplitude and phase accuracy between arrays. Studies show that although the SPIR algorithm solves the blind zone problem well, the performance stability is affected greatly by ground clutter, because the ground clutter results in the increase of condition number of Point Spread Function (PSF) matrix, so clean algorithm must be incorporated to overcome the problem. Simulations show that as the element number of array increases, the ability of putting up with the above errors of SPIR algorithm increases correspondingly. The accurate measurement of real position is much more important than the control of position for a designed distributed architecture in some degree.Chapter 6 gives reviews and concluding remarks for the whole work in the dissertation, and at the same time points out some limitations and proposes four research aspects, which should be aimed to solve in the future.更多还原