时-空联合压缩感知3D超分辨雷达成像机理研究

Super-resolution 3D imaging radar has widely applications in aeras such as medical detection, through-wall seeing, life-sign exploration, archeological investigation and civil engineering，etc. However, Super-resolution imaging radar based on the matched filter and UWB technology is facing big challenges in aspects of high rate sampling, huge amount of storage and complexity of signal processing.Compressive sensing(CS) provides a new way to solve bottlenecks faced by super-resolution radar imaging of small sparse scene.It is quite different from the matched filtering，introduces a new concept based on the direct convertion from analog signal to information, and can recovery the original signal at high probability by the reconstructing algorithm with few non-relevant measurements sampled at a rate significantly below the Nyquist rate. Aiming at the tough problem of 3D super-resolution imaging of the medical detection,a new system of a virtual cylindrical surface synthetic aperture based on MIMO techniques and circular SAR and a scheme of target signal acquisition and processing based on the joint time-space compressive sensing are proposed in this project,in which new ideas on modern radar design and compressive sensing are combined. The goal of the project is to build a theoretical system framework of MIMO-CSAR system after researching the basic principles of the 3D super-resolution imaging based on TSCS. Researches will focus on key problems about the optimal design of MIMO-CSAR ,the sparse representaion of MIMO-CSAR return signals,non-relavant measurement,and reconstructing algorithm of target scene.It will lay a solid theoretical foundation for solving current bottlenecks on super-resolution radar imaging and exploring new theory on the 3D radar imaging based on TSCS.

超分辨3D成像雷达在医疗诊断、隔墙透视、生命探测、考古发掘和民用工程等方面具有广泛的应用,但基于传统匹配滤波理论的超高分辨UWB成像雷达正面临着采样速率高、海量存储和信号处理复杂等挑战性问题。压缩感知不同于传统匹配滤波，能以远低于Nyquist采样率获取少数非相干测量数据，通过重构算法高概率恢复信号，给解决高分辨雷达成像的瓶颈问题的提供了新的途径。本课题针对3D超高分辨医疗诊断雷达成像的科学问题，将现代雷达设计新理念与压缩感知理论相结合，提出了基于MIMO和圆周SAR相结合的虚拟柱面合成孔径系统和基于MIMO-CSAR的联合时-空压缩感知（TSCS）目标信号获取与处理方法；研究基于TSCS的3D超分辨成像机理,构建MIMO-CSAR系统理论架构；重点解决MIMO-CSAR系统优化设计和基于TSCS的MIMO-CSAR 3D目标信号的稀疏表示、联合时-空非相干测量和3D目标场景重构问题。

首先，根据超宽带层析成像雷达医学检测的需要，研究了人体组织和器官电磁散射特性，建立了人体器官分层简化仿真模型和网格化小场景目标模型。研究了基于时-空压缩感知的MIMO-CSAR成像理论，建立了小场景医学3D探测系统实现架构，完成了小场景目标MIMO层析成像实验验证。. 其次，针对高分辨MIMO雷达成像在高速率采样、海量数据和实时性等方面等信号处理基本问题，重点研究了基于压缩感知的MIMO-CSAR信号处理的关键技术。提出了基于Costas-LFM脉冲和基于Morlet小波集合的UWB MIMO-CSAR雷达时空正交波形设计方法。并在此基础上，研究了这两种特定波形字典对应的压缩投影变换方法，提出了基于Stretch处理和基于EMD的Morlet小波信号分解压缩变换处理方法。并进一步研究了基于信息理论和基于信杂比最优化的自适应波形设计方法。. 接着，根据医学小场景检测的实际情况，研究了基于CS的MIMO-CSAR成像雷达目标信号的非相干测量方法，建立了TSCS的测量网格化测量的数学模型；研究了基于多路并行处理技术的虚拟阵列信号处理及其外推技术，提出了基于虚拟阵与分数延时滤波器的高分辨力超宽带波束形成新方法。.随后，研究了基于传统l0重构方法进行MIMO-CSAR雷达层析成像算法，证明用OMP方法系统成像的可行性。提出了基于Morlet小波集和基于EMD算法的MIMO-CSAR成像新方法。研究了基于UWB MIMO-CSAR雷达成像系统的优化方法。研究了基于时间逆处理的UWB雷达成像方法，提出了基于SF-DORT和基于SF-MUSIC的成像改进方法。. 最后，研制了用于UWB MIMO雷达层析成像实验的UWB天线系统。完成了基于LFM和Stretch处理的雷达验证实验。构建了用于医疗诊断的UWB MIMO雷达仿真演示系统。