零辅助数据MIMO雷达自适应检测问题研究

项目名称： 零辅助数据MIMO雷达自适应检测问题研究

项目编号： No.61501351

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 无线电电子学、电信技术

项目作者： 刘军

作者单位： 西安电子科技大学

项目金额： 21万元

中文摘要： MIMO雷达是近年来广受关注的一种新体制雷达，其空间分集和信号分集可以用来缓解目标RCS闪烁和抑制杂波，在目标检测方面具有较明显优势。传统的MIMO雷达自适应检测算法通常假设存在大量均匀的辅助数据用来估计杂波协方差矩阵，这对辅助数据的均匀性和充足性提出了非常苛刻的要求，从而使得传统方法的性能在非均匀杂波环境中下降明显。为了消除因辅助数据匮乏和不均匀而引起的性能下降，本项目拟提出不需要辅助数据的MIMO雷达自适应检测算法，具体研究内容包括：（1）分布式MIMO雷达零辅助数据自适应检测算法的设计；（2）集中式MIMO雷达零辅助数据自适应检测算法的设计及理论分析；（3）基于检测概率最大化准则的零辅助数据MIMO雷达波形优化。通过上述研究，将会丰富MIMO雷达自适应检测理论，拓宽MIMO雷达的应用范围，以及增强MIMO雷达的探测能力。因此，本研究具有重要的理论意义和应用价值。

中文关键词： MIMO雷达；自适应检测；波形设计；辅助数据

英文摘要： MIMO radar has recently received intense interest as a new technology. It can be used to reduce the scintillations of targets' RCS and suppress clutter, and hence has an obvious superiority in target detection. It is usually assumed in the traditional MIMO radar adaptive detection algorithms that a large number of homogeneous training data are available for clutter covariance matrix estimation. This assumption has demanding requirements on the homogeneity and sufficiency of the training data, which leads to an obvious loss in detection performance of the traditional algorithms in non-homogeneous environments. In order to avoid the performance loss caused by the deficiency and non-homogeneity of the training data, we propose adaptive detection algorithms without training data in MIMO radar, which include the following contents: (1) design adaptive detection algorithms without training data in distributed MIMO radar; (2) design and analyze adaptive detection algorithms without training data in co-located MIMO radar; (3) MIMO radar waveform optimization without training data according to the criterion of maximizing detection probability. These investigations would enrich adaptive detection theory in MIMO radar, expand the application scope of MIMO radar, and improve the detection capability of MIMO radar. Therefore, these studies have important theoretical significance and application value.

英文关键词： MIMO radar;adaptive detection;waveform design;training data