基于空防雷达网络的多传感器信息融合关键技术研究及其应用

发布时间：2018-11-18 14:44

【摘要】：近年来,海湾战争及911恐怖袭击、马航MH370失联等国际事件暴露出的空防情报不共享、提供目标属性不明确、航迹不连续、态势评估不准确等问题,严重影响了战争行动和反恐作战。我军受指挥关系、武器性能、地理遮蔽、编制体制等条件约束,不同程度地存在着上述问题,因此,从2010年开始,重点加强了对空防雷达网络情报系统的研究和建设。本论文所研究的课题,系作者从事几十年雷达情报系统建设与保障维护工作基础上,结合工作实践经验所做的创新和探索。课题重点研究了空防雷达网络系统在多传感器信息融合中的优化布站、目标识别、目标跟踪、威胁评估等关键技术,探讨了有关技术的改进问题,并在工作实际中进行了应用。课题的研究较好地发挥了多传感器信息融合技术在空防雷达网络中的作用,提高了空防雷达网络系统整体作战性能。论文以实现空防雷达网络系统的布站优化,改进异类传感器信息融合的目标识别,完成被动传感器的目标跟踪,以及分析空防威胁评估为目标,主要基于蚁群算法(ant colony optimization, ACO)、模糊神经网络、高斯-厄密特滤波(Gauss-Hermite filtering, GHF)、贝叶斯网络(Bayesian networks, BN)等信息融合理论及方法,对空防雷达网络系统的布站优化、目标识别、目标跟踪、威胁评估等展开研究及实践应用。论文首先归纳了多传感器信息融合的特点和基本定义。描述了信息融合的系统模型和主要算法,异类传感器系统可以实现信息互补,避开各自缺点,但融合复杂,没有统一方法：被动传感器本身不发射信号,具有较强的生存能力和反隐身能力。其次,基于雷达网络布站方案及优化方法,实地搭建了空防雷达网络系统平台。研究了基于ACO算法的雷达网络布站方案,提出了基于参数优选演进的优化布站方法,仿真得到了较好的布站方案；进行了实地勘察,完成了空防雷达网络系统的实地搭建,验证了理论方法的正确性和可操作性。再次,研究了空防雷达网络系统基于异类传感器信息融合的目标识别改进问题。讨论了基于3种异类传感器构建异类传感器信息融合系统,研究了特征层基于模糊神经网络的改进聚类中心选择方法和决策层改进D-S证据组合矛盾冲突解决方案；对四批四架飞机日标进行了侦测,得到了目标特征参数,进行了仿真实验,完成了空防雷达网络系统对空中目标融合识别的改善。之后,基于多被动传感器系统,讨论了空防雷达网络纯方位多目标跟踪融合方法。分析了被动雷达纯方位条件下基于高斯-厄密特滤波(Gauss-Hermite filtering, GHF)的目标航迹关联更新方法,利用目标先验信息计算得到全体观测与目标的可行划分；基于空防雷达网络系统,对两批两架飞机目标进行了纯方位被动跟踪关联,并对关联结果进行了比较。最后,结合空防雷达网络系统对多种情况下的敌机威胁等级进行了评估分析。给出了威胁评估的数学描述,建立了模糊动态贝叶斯网络(fuzzy dynamic Bayesian networks, FDBN)模型,对两批两架飞机目标进行了模拟空战的雷达侦测,并通过计算机仿真给出了基于模糊动态贝叶斯网络的威胁评估(threat assessment using fuzzy dynamic Bayesian networks, TAFDBN)推理过程,以及FDBN灵敏度计算,研究结果扩展了空防雷达网络系统的作战应用。
[Abstract]:In recent years, the air-raid information that has been exposed by international events such as the Gulf War and the September 11 terrorist attacks and the Malaysia Airlines flight MH370 is not shared, and the problems such as the unclear target attribute, the non-continuity of the track, the inaccurate situation assessment and the like are provided, and the war operation and the counter-terrorism operation are seriously affected. As a result, the research and construction of the airborne anti-radar network information system has been strengthened in the beginning of 2010. The subject of this thesis is the innovation and exploration of the author's work experience in the construction and maintenance of radar information system for several decades. This paper focuses on the key technologies such as the optimization of the air defense radar network system in the multi-sensor information fusion, the target recognition, the target tracking and the threat assessment, and discusses the improvement of the related technologies and the application in the practice. The research on the problem of multi-sensor information fusion plays an important role in the anti-radar network, and the overall operational performance of the airborne radar network system is improved. Based on ant colony optimization (ACO) and fuzzy neural network, the paper is based on ant colony optimization (ACO) and fuzzy neural network. The information fusion theory and method of Gauss-Hermite filtering (GHF) and Bayesian networks (BN) are used to research and practice the optimization, target recognition, target tracking and threat assessment of the airborne radar network system. The paper first summarizes the characteristics and the basic definition of multi-sensor information fusion. The system model and the main algorithm of the information fusion are described. The heterogeneous sensor system can realize the information complementation and avoid the respective disadvantages, but the fusion is complex and there is no uniform method: the passive sensor does not transmit the signal itself, and has strong survivability and anti-stealth capability. Secondly, the airborne radar network system platform is built on the ground based on the radar network fabric station scheme and the optimization method. This paper studies the radar network fabric station scheme based on the ACO algorithm, and puts forward the optimized cloth station method based on the parameter optimization evolution. The simulation results in the better cloth station scheme. The field investigation is carried out, and the field construction of the air defense radar network system is completed. and the correctness and the operability of the theoretical method are verified. Thirdly, the target recognition improvement problem based on the heterogeneous sensor information fusion is studied. This paper discusses the construction of heterogeneous sensor information fusion system based on three kinds of heterogeneous sensors, and studies the improved clustering center selection method and decision-making level improvement of D-S evidence combination conflict resolution solution based on the fuzzy neural network. The target characteristic parameters are obtained, and the simulation experiment is carried out, and the improvement of the air target fusion recognition by the air defense radar network system is completed. Then, based on the multi-passive sensor system, the method of multi-target tracking and fusion of the pure azimuth of the airborne radar is discussed. The target track correlation updating method based on the Gaussian-Hermite filtering (GHF) under the pure azimuth condition of the passive radar is analyzed, and the feasible division of the whole observation and the target is obtained by using the prior information of the target, and the method is based on the air defense radar network system, A pure-orientation passive tracking association for two aircraft targets is carried out, and the correlation results are compared. Finally, the threat level of the enemy aircraft under various conditions is analyzed and analyzed in combination with the airborne radar network system. The mathematical description of the threat assessment is given, and the fuzzy dynamic Bayesian network (FDBN) model is established, and the radar detection of the two aircraft targets is simulated. The reasoning process of the threat assessment (TAFDBN) based on the fuzzy dynamic Bayesian network and the sensitivity calculation of the FDBN are given by computer simulation. The results of the research extend the operational applications of the airborne radar network system.
【学位授予单位】：东北大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN959;TP202

【相似文献】