基于连续域稀疏特征的超分辨MIMO雷达定位技术研究

In the problem of target location with multi-input and multi-output (MIMO) radar, the super-resolution location algorithms need a large number of snapshots and have high complexity. The amount of data can be reduced by the compressed sensing theory when the features of target sparsity are exploited, but the discretized grids in spatial spectrum result in the basis mismatch error. Therefore, this project takes the MIMO radar as the research object, and establishes a super-resolution location system based on the target sparse features on the continuous domain: A sparse reconstruction model based on the continuous domain will be established, and the joint approximate message passing (AMP) and atomic norm algorithm will be studied to localize the target; The theoretical performance of sparse reconstruction will be studied based on the state evolution, and the target localization performance of super-resolution MIMO radar will be theoretically deduced; Combining with the sparse reconstruction algorithm on the continuous domain, the fast optimization algorithm based on alternating direction method of multipliers (ADMM) will be studied in the MIMO radar. The following key scientific problems will be focused: with the limit grid refinement (continuous domain), the establishment and solution of the sparse reconstruction model under the condition of strong mutual coherence; the theoretical description of sparse location performance in the super-resolution MIMO radar; the system optimization combining with the sparse reconstruction algorithm, et al. Some innovative research results will be obtained to provide an important theoretical basis for the development of radar location technology.

在多输入多输出（MIMO）雷达目标定位问题中，超分辨定位算法需要大量快拍数且复杂度高，通过挖掘目标稀疏特征的压缩感知理论可降低数据量，但空间谱的离散网格化会导致基失配误差，限制了MIMO雷达定位性能的进一步提升。为此，本项目拟以MIMO雷达为研究对象，建立一套基于连续域稀疏特征目标的超分辨定位系统：构建基于连续域的稀疏重构模型，并研究基于近似信息传递（AMP）与原子范数相融合的定位算法；研究基于状态演化的稀疏重构性能分析方法，理论推导超分辨MIMO雷达的目标定位性能；联合连续域稀疏重构算法，研究基于交替方向乘子（ADMM）的MIMO雷达快速优化算法。重点研究所遇到的关键科学问题：网格极限细化（连续域）时强相干条件下，稀疏重构模型的建立与求解，超分辨MIMO雷达稀疏定位性能的理论描述以及联合稀疏重构算法的系统优化等。取得一些创新性的研究成果，以期为雷达定位技术的发展提供重要理论依据。

在MIMO（Multiple-Input and Multiple-Output）雷达目标定位问题中，超分辨定位算法需要大量快拍数且复杂度高，通过挖掘目标稀疏特征的压缩感知理论可降低数据量，但空间谱的离散网格化会导致基失配误差，限制了MIMO雷达定位性能的进一步提升。为此，本项目以MIMO雷达为研究对象，建立了一套基于连续域稀疏特征目标的超分辨定位系统：构建了基于连续域的稀疏重构模型，并研究了基于近似信息传递与原子范数相融合的定位算法；研究了基于状态演化的稀疏重构性能分析方法，理论推导了超分辨MIMO雷达的目标定位性能；联合连续域稀疏重构算法，研究了基于交替方向乘子法（Alternating Direction Method of Multipliers, ADMM）的MIMO雷达快速优化算法。依托本项目发表了23篇SCI或EI检索的高质量期刊论文，其中，SCI论文22篇，IEEE期刊论文11篇，中科院分区顶级Top期刊论文7篇，ESI高被引论文1篇；申请了29项国家发明专利，其中，3项已经授权；在人才培养方面，培养了1名青年教师、4名博士研究生和4名硕士研究生；主要科研成果获得了江苏省科技奖（三等奖）。