辅助信息多秩子空间信号检测方法研究

Multi-channel array adaptive detection has received considered attention since it is adaptive to the character of unknown clutter spectrum. However, it is difficult to collect adequate homogeneous training samples in the complex and multivariate scenarios where radar works. This will degrade the detection performance greatly. In addition, the mismatch of the steering vector will affect the robustness of detectors. . This project focuses on the effective and robust detection for multi-rank subspace target in terms of the exploiting of the persymmetric covariance structure information, symmetric power spectrum information and the lower order time related structure information of clutter in the background of homogeneous and partially homogeneous environments. This project can provide the new theories and methodologies for knowledge-aided detection for multi-rank subspace signal. The main contents of the study include: Establishing the homogeneous and partially homogeneous signal model for multi-rank subspace target; Studying knowledge-aided method for GLRT test, Rao test and Wald test; Designing the corresponding subspace signal detectors based on the one step procedure and two-step procedure; Analyzing and comparing the performance of the proposed detectors. The aim of this project is to improve the detection performance for subspace target and to overcome the technical difficulties of deficient training samples and the steering vector mismatch in practical application.

多通道阵列自适应信号检测因其对未知杂波谱特性具有自适应性而成为雷达目标探测领域的研究热点。然而雷达工作环境的复杂多变性会导致实际中难以获得足够的均匀训练样本，从而致使检测性能急剧下降。此外，导向矢量失配也会影响检测器的稳健性。. 本项目以均匀和非均匀杂波为背景，拟从辅助反对称协方差矩阵结构信息、对称功率谱信息及杂波低阶时间相关性结构信息三个方面来实现有效稳健的多秩子空间目标检测，建立辅助信息多秩子空间信号检测的新理论新方法。拟研究的主要内容包括：针对多秩子空间目标，建立均匀及部分均匀的信号模型，研究有效利用辅助信息的GLRT检验、Rao检验和Wald检验方法，基于一步法和两步法策略，设计相应的子空间目标自适应检测器，并在此基础上分析及比较检测性能，提高子空间目标的检测性能，突破实际应用中训练样本数不足、导向矢量失配等技术难点。

项目针对雷达工作环境的复杂多变性会导致实际中难以获得足够的均匀训练样本，从而致使检测性能急剧下降以及导向矢量失配也会影响检测器的稳健性。.本项目以均匀和非均匀杂波为背景，辅助协方差矩阵结构信息来实现有效稳健的多秩子空间目标检测，建立辅助信息多秩子空间信号检测的新理论新方法，提高了子空间目标的检测性能，突破实际应用中训练样本数不足、导向矢量失配等技术难点。自2018年1月立项以来，项目在研期间，课题组先后共计发表高质量学术论文13篇，其中SCI索引8篇,EI索引9篇，已标注项目资助论文10篇，其中第一标注8篇，总标注率达到77%；授权专利1项，受理专利8项；多次参加国内相关领域知名学术研讨会，并先后于会议期间做主题及学术报告3场；项目在研期间，前后累计培养博士研究生3人（其中协助课题组组长培养两名博士生，均已毕业），硕士研究生4人。截至2020年11月底，本课题的全部研究内容已经完成，达到并部分超出合同规定的研究成果指标。