基于卫星导航信号的相异多孔径无源成像技术

The passive imaging radar using navigation satellite as illuminators of opportunity is an important branch of modern radar technology due to its particular advantages, such as large covers areas, short revisit period and good time synchronization. A multi-aperture passive imaging radar system, which consists of GNSS illuminators and stationary ground-based receivers, is studied in the wavenumber domain. The Fourier transform relationship between target scattering function in the space domain and the echo in the wavenumber domain is proved, then it can be derived that the imaging resolution depends on the support region of the echo. The distribution of echo in the wavenumber domain is jointly determined by both the bandwidth of transmitted signal and the size of the synthetic aperture, and it is difficult to identify the difference between their contributions to the resolution in the dissimilar multi-aperture system. Therefore, it is possible to achieve a high resolution image using narrow-band illuminators by improving the number and size of the aperture. The adjusting of the position of receivers is the only way to optimize the distribution of echo in the wavenumber domain, because other factors, such as satellite orbit and the transmitted signal, are uncontrollable for the passive system. Thus, it is one of the key contents in our research. Taking into account the ultra-narrow bandwidth of the GNSS signal and the uncontrollable configuration of the illuminators, the echoes constitute a non-uniform and discontinuous coverage in wavenumber domain, which leads to poor imaging performance. Two methods, compressed sensing and scatter-by-scatter matched filtering, are used to reconstruct the images. Finally, the comprehensive utilization of all GNSS systems including GSP, GALILEO, GLONASS and COMPASS to improve the resolution will be researched.

基于卫星导航信号(GNSS)的无源成像技术具有多种突出优点，已成为微波成像的重要发展方向。本项目以波数域回波与空域目标像之间的傅里叶变换关系为出发点，借助波数域分析手段开展基于GNSS卫星照射、地面固定站接收的相异多孔径无源成像技术的研究，着重解决GNSS信号带宽过窄制约系统成像分辨率和孔径差异过大不利于成像反演两个难题。在理论分析层面，研究波数域回波样本分布形式及支撑域尺寸与成像分辨率之间的内在联系，突破距离向分辨率仅依赖于信号带宽、方位向分辨率仅依赖于成像孔径的传统界线，实现两者在波数域的统一，同时寻求两者的互补途径；在技术突破层面，针对相异孔径造成的波数域样本支撑域不连续问题，展开基于压缩感知和逐点匹配滤波的图像反演方法的研究；在应用实施层面，研究多种类多频段GNSS照射源的综合利用方法，建立以提高成像分辨率为目的的接收站位置调整策略，最终建立基于GNSS照射的多通道无源成像方案。

全球卫星导航系统(Global Navigation Satellite System, GNSS)的卫星数量多、覆盖范围广、重访周期短，与地面接收站构成的无源成像系统具有抗干扰能力强、几何构型灵活、成本低、可连续观测等多种突出优点，因此很适合用于高危滑坡体、冰川、积雪的成像监测。本项目围绕带宽与孔径对成像分辨率贡献互补和波数域采样数据大面积缺失条件下的图像反演两大科学问题，研究了波数域回波样本分布形式及支撑域尺寸与成像分辨率之间的内在联系，利用带宽和孔径在波数域的统一描述，利用孔径的扩展补充GNSS信号带宽的不足，提高系统成像分辨率。推导了基于GNSS信号的无源成像系统平均广义模糊函数，消除了码元序列伪随机性对模糊函数的影响，通过非相干融合和相干融合两种方式推导了多发多收系统的模糊函数。以分辨椭圆面积最小为目标函数，开展了接收站布站优化研究，初步建立了布站规则。项目还开展了稀疏重构方法对接收机布站的约束研究，分析了系统观测矩阵中的互相关系数参量与收发站布局之间的内在联系，推导了最小化互相关系数准则对接收机布站位置优化的约束。利用递推方法实现对不断接收GNSS信号的序贯处理，实提出了序贯匹配滤波成像和序贯伪逆成像两种成像方法。探索了基于分层贝叶斯先验的稀疏成像、基于改进的原子范数最小化的多通道无源雷达高分辨成像、基于贝叶斯压缩感知的稀疏自聚焦成像、结合总变差正则化约束的改进稀疏成像等多种稀疏成像方法，最终将基于非凸函数约束的优化成像算法、基于Laplace先验的稀疏贝叶斯学习高分辨成像、基于双曲正切约束的稀疏成像三种算法作为主要的稀疏图像反演重构方法。从模糊函数角度研究了不同导航卫星信号和不同的导航信号组合对系统分辨性能的影响，给出多种卫星导航系统下的布站准则，建立了基于多颗卫星照射、单个地面固定接收站的系统方案。