组网雷达动态规划检测前跟踪方法研究

This project mainly deals with the fundamental and theoretical problems of dynamic programming based track before detect method for netted radar systems (NRS-DP-TBD). The NRS-DP-TBD can be established by studying the following fundamental problems: the normalized likelihood function of NRS measurements, the integration and fusion theory of the multiple frame measurements obtained by NRS and the target detection method after the DP-TBD based integration and fusion. We will solve the existing key problems, and propose a complete NRS-DP-TBD signal processing framework. In addition to the theoretical studies, all the new models and algorithms proposed in this project will be tested by a scaling experiment which is composed of 3 to 5 small radars. This project is of great significance, since it can provide the basic models, integration and fusion criteria and signal processing algorithms for the applications of TBD techniques to multi-static radar systems, e.g., NRS. Moreover, the theoretical contributions can also be utilized when applying DP-TBD to other multi-sensor systems, such as target detection and tracking in sonar network system and infrared/optical sensor networks.

针对组网雷达动态规划检测前跟踪方法存在的一系列模型和理论问题，本项目拟开展组网雷达量测的标准似然函数建模方法、多雷达多周期量测的积累融合方法、积累融合后值函数域的目标检测算法等研究内容，建立和完善组网雷达动态规划检测前跟踪理论体系；突破关键技术，提出一套组网雷达动态规划检测前跟踪信号处理算法，并利用3-5台小型雷达搭建原理验证实验平台，采集实测数据，对本项目研究所获得的模型和算法成果进行验证。该项目研究可以为检测前跟踪技术推广应用到组网雷达等多站雷达系统中奠定理论和技术基础，具有重要的科学意义；其理论和技术成果还可应用于组网声纳探测、红外/光学多传感器目标检测等领域。

组网雷达系统通过节点间协同工作显著提高系统探测能力。当目标信噪比低或起伏严重时，常规的组网雷达系统中采用的先检测后跟踪方式会造成目标漏检，导致系统性能严重受损。针对这一问题，本项目主要围绕组网雷达动态规划检测前跟踪方法存在的一系列模型和理论问题，开展了组网雷达系统量测标准似然函数建模、多雷达多周期量测积累融合方法、值函数域的检测算法与性能分析等三个方面的研究，形成了一系列创新性研究成果，主要包括：基于后验信息融合和基于点迹序列融合的NRS-MF-TBD算法、基于多假设检验的无延迟检测判决算法、非均匀背景杂波值函数域的自适应检测门限算法等方法。相关成果共发表学术论文30篇，其中在IEEE T-SP和IEEE T-AES等高水平国际学术期刊上发表18篇；在雷达会议、融合会议等国际会议发表学术论文11篇；申请专利6项，授权专利16项。本项目的研究成果提高了组网雷达系统微弱多目标探测的基础研究水平，可应用于组网雷达系统、组网声呐系统、红外/光学多传感器系统等，具有重要的科学意义和应用价值。