基于分布式压缩感知的MIMO雷达弱小目标定位与跟踪方法研究

项目名称： 基于分布式压缩感知的MIMO雷达弱小目标定位与跟踪方法研究

项目编号： No.61501355

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： 李彩彩

作者单位： 西安电子科技大学

项目金额： 19万元

中文摘要： 目标雷达截面积的角闪烁以及现代战争中隐身目标的出现增加了传统相控阵雷达发现目标的难度。分布式MIMO雷达由于多个通道从不同的角度观测目标，能够有效抑制雷达RCS闪烁，提高隐身目标的探测性能，但同时也面临更大的数据量和计算量挑战。压缩感知理论通过开发信号的稀疏特性，在远小于Nyquist采样率的条件下对信号进行采样及重建，能够获得数据量与计算能力的改善。充分利用雷达目标场景的稀疏特性，将压缩感知理论应用到MIMO雷达信号处理，将有效减轻数据存储，传输及计算的负担。而目标定位与跟踪是雷达系统的基本任务，因此本项目将针对分布式MIMO雷达系统，在稀疏建模框架下，研究弱小目标的定位与跟踪问题，主要内容包括：（1）复杂环境下基于分布式压缩感知的MIMO雷达稀疏建模及恢复；（2）建立准确的目标量测模型及动态模型；（3）基于PHD滤波的检测前跟踪算法。

中文关键词： 分布式MIMO雷达；分布式压缩感知；稀疏回复；检测前跟踪；概率假设密度

英文摘要： The scintillation of the target’s radar cross section and the emergence of the stealth targets in modern war increase the difficulty of conventional phased array radar to find a target. Since the distributed MIMO radar observations a target from different angles, it can effectively suppress the RCS scintillation and improve the detection performance of stealth target, while faces a greater amount of data and computation challenge. Compressive sensing allows us to accurately reconstruct data from significantly fewer samples than the Nyquist rate if the received signal is sparse in some basis representation. Since the number of targets in a radar scene is often limited, we can introduce compressive sensing theory to MIMO radar signal processing, so as to effectively reduce the data storage, transmission and computing burden. The target location and tracking is the basic task of the radar system, therefore this project will focus on the multiple weak target location and tracking problem of the distributed MIMO radar system. The research is conducted in the framework of sparse modeling. The main.content includes: (1) DCS based sparse modeling and recovery of the distributed MIMO radar under complex environments. (2) Establish a high fidelity measurement model and dynamic model. (3) Track before detection technique based on PDH filter.

英文关键词： distributed MIMO radar;distributed compressive sensing;sparse recovery;track-before-detect;probability hypothesis density filter