基于多目标进化算法的稀疏随机步进频全极化雷达空间目标成像

Fully polarimetric ISAR imaging is of great significance for space target recognition. Sparse random stepped frequency (SRSF) signals possess the merits such as narrow instantaneous bandwidth and easy-to-generation. For the fully polarimetric radar with SRSF waveforms, high resolution fully polarimetric ISAR imaging can be achieved via compressive sensing. However, the sparse imaging methods based on predesigned dictionary suffers or even fail due to the unknown target motion parameters, while the existing motion compensation and cross-range scaling methods for compressive sensing ISAR are mainly developed for single polarisation radar, with fully polarimetric information unefficient exploited. This proposal presents a novel scheme for sparse imaging. In this scheme, firstly, a joint sparse representation model including motion parameters is established for fully polarimetric radar signals. Then, a multiobjective optimization frame including sparse optimization and optimization for find an optimal polarimetric channel is built, and an approach for optimizing simultaneously these two kinds of objectives based on evolutionary multiobjective algorithm. Parametric sparse recovery exploiting jointly sparse prior and fully polarimetric information can be achieved by resorting to this scheme. Based on this present scheme, two main research contents “Synthetic high resolution range profile (HRRP) motion compensation for SRSF fully polarimetric radar” and “Joint fully polarimetric ISAR imaging and cross-range scaling via compressive sensing” is planned to implemented so as to solve the problems of motion compensation and cross-range scaling for SRSF fully polarimetric ISAR imaging, urging the progress of the fully polarimetric high-resolution imaging for the space target.

全极化ISAR成像对于空间目标识别具有重要意义。稀疏随机步进频（SRSF）信号具有瞬时带宽小、易于实现等优点，与全极化雷达结合可实现全极化高分辨压缩感知ISAR成像。然而，目标运动参数未知使常用的基于预先设计词典的稀疏成像方法性能下降甚至失效，现有压缩感知ISAR运动补偿和方位定标方法主要针对单极化雷达，不能有效利用全极化信息。本项目提出“构建全极化雷达信号含运动参数的联合稀疏表示模型，将稀疏优化和最优极化选取纳入统一的多目标优化框架，并以多目标进化算法为基础设计同时优化这两类目标函数的方法，实现联合利用稀疏先验和全极化信息的参数化稀疏恢复”这一稀疏成像的崭新思路。基于该思路，通过研究“SRSF全极化雷达合成高分辨距离像（HRRP）运动补偿”和“全极化压缩感知ISAR联合自聚焦成像与方位定标”，解决SRSF全极化压缩感知ISAR运动补偿和方位定标难题，推动空间目标全极化高分辨成像的进步。

雷达高分辨成像在探测与识别领域中具有重要价值。本项目组以空间攻防对抗为研究背景，围绕运动目标全极化高分辨信息获取这一研究目标，重点突破了“如何综合利用目标的稀疏先验信息、全极化信息等多种信息提高运动补偿精度和成像质量”这一问题，取得的成果主要包括：. 建立了全极化稀疏随机步进频雷达信号模型、提出了全极化一维距离像合成的多目标优化方法；建立了全极化ISAR方位信号参数化模型，提出了全极化ISAR联合自聚焦与方位定标的多目标优化方法；提出了压缩感知和稀疏恢复成像算法的性能评估指标体系、性能评估方法；在已有研究基础上，将所提多目标优化算法与空间信息网络体系结构设计工作交叉融合，额外完成了多目标优化算法在空间信息网络体系架构设计方面的拓展应用，提出了基于多目标优化的临近空间段高空平台（HAP）通信网络、卫星通信网络等空间信息网络的体系结构设计方法。同时，在极化雷达成像试验设计过程中，在本项目的资助下开展了部分天线与射频系统设计工作。. 研究成果对雷达目标探测与识别领域的发展具有一定的促进作用，特别是，本项目为多目标优化在雷达和通信领域的创新应用做出了有益探索。以项目研究成果为基础，出版译著1部，申请发明专利3项（1项已授权），发表学术论文13篇，其中在IEEE期刊发表论文5篇，SCI检索8篇。