基于复杂电磁大数据的辐射源联合参数模型挖掘和识别

With the rapid development of sensor technology, numerous new emitter types have emerged continuously in reconnaissance, surveillance, orientation, tracing and navigation in the air, sea, and land. With the complex electromagnetic big data，one typical characteristic of these new emitter types is that their signal parameters overlap heavily and exhibite some kinds of complex joint regulation patterns. It has been such a great challenge to identify these new emitter accurately in real time. It has become almost impractical to identify new emitter types and their working modes without making use of their joint regulation patterns. However, due to the incompleteness and uncertainty of emitter data and the complexity of joint parameter regulation patterns, the joint parameter regulation patterns are difficult to analyze, discover or use. In this project, we aim to study how to discover the joint-parameter patterns incrementally from the dynamically increasing and incomplete complex electromagnetic big data. Based on that, we futher study to build a emitter classification model with the discovered joint-parameter patterns to enhance new emitter identification. Our study is helpful to improve the accuracy, robustness and timeliness of new emitter identification in various critical application scenarios, i.e., aviation, navigation, salvage, traffic control and electronic countermeasure.

随着传感器技术的发展，在陆、海、空侦察、监视、定位和跟踪等应用领域涌现出越来越多种类的新体制辐射源。在当前复杂电磁大数据环境下，各类新体制辐射源信号特征参数相互重叠且具有联合变化特点和模式。新体制辐射源的实时性和准确性识别是一项艰巨的挑战。如果不掌握其联合变化规律，仅仅依靠辐射源常规统计参数特征，已难以区分辐射源类型。然而，由于新体制辐射源信号参数联合变化模式非常复杂，信号参数范围重叠严重，采集到的辐射源信号数据不完整、不确定性高，信号参数联合变化模式难以分析、挖掘和利用。本课题拟采用增量式学习的辐射源联合参数建模方法，从海量递增的、不完整、不确定性高的复杂电磁大数据中挖掘出典型的辐射源联合参数变化模式，并在此基础上构建识别模型，以提高新体制辐射源的识别能力。本课题的研究成果对于复杂电磁大数据环境下提升航空、航海、导航、搜救、交管和电子对抗中辐射源识别的准确性、鲁棒性和实时性具有重大意义。

随着传感器技术的发展，在陆、海、空侦察、监视、定位和跟踪等应用领域涌现出越来越多种类的新体制辐射源。在当前复杂电磁大数据环境下，各类新体制辐射源信号特征参数相互重叠且具有联合变化特点和模式。新体制辐射源的实时性和准确性识别是一项艰巨的挑战。如果不掌握其联合变化规律，仅仅依靠辐射源常规统计参数特征，已难以区分新体制辐射源类型。本课题采用了增量式学习的辐射源联合参数建模方法，从海量递增的、不完整、不确定性高的复杂电磁大数据中挖掘出典型的辐射源联合参数变化模式，并在此基础上构建识别模型，以提高新体制辐射源的识别能力。.在雷达辐射源联合参数建模方面，本项目采用了联合参数模型描述新体制辐射源的联合变化规律，体现了辐射源各信号特征参数变化的联合性、协同性。该联合参数模型既考虑了标量型特征、区间型特征、随机模式型特征，也考虑了复杂的联合型参数特征。.在雷达辐射源信号参数识别及评估方面，本课题针对不完整性、不确定性程度高的复杂辐射源信号参数，开展了辐射源联合参数识别模型构建研究，分别提出了针对标量型、区间型、随机模式型与混合型等复杂辐射源信号参数识别模型，并针对海量辐射源信号参数数据构建了基于极限学习机的辐射源信号参数识别模型。.在面向辐射源识别的联合参数增量式学习及评估验证方面，本课题针对标量型、区间型、随机模式型和联合性特征参数分别提出了增量式学习方法，并结合仿真和实测数据进行了评估验证。此外，本课题结合协同探测管控和信号子空间预测控制开展了模型方法的应用验证。.本课题的研究成果对于复杂电磁大数据环境下提升航空、航海、导航、搜救、交管和电子对抗中辐射源识别的准确性、鲁棒性和实时性具有重大意义。