基于信号特征知识图谱和特征融合的特定辐射源识别技术

Specific Emitter Identification Technology Based on Signal Feature Knowledge Graph and Feature Fusion

SEI可以用于识别各种类型的无线电发射源,包括无线电通信设备、雷达系统、无线电干扰源等。现有的SEI方法往往通过深度学习来实现,通常SEI要求分类模型具有很高的准确性和鲁棒性。针对现今辐射源信号可视化管理不足和信号识别准确率低的问题,提出了一种信号特征KG和特征融合的SEI方法。第一,该方法创新性地建立了信号特征数据库,且利用KG对信号特征实现可视化表征;第二,基于构建的信号特征KG进行特征融合,有效提升了特定辐射源分类识别的准确率。仿真结果表明,所提出的基于KG和特征融合的SEI方法可以更好地对辐射源信号进行可视化管理,且提升了辐射源识别性能。

Specific emitter identification(SEI)can be used to identify various types of radio emission sources,including radio communication equipment,radar systems,radio interference sources,etc.Existing SEI methods are often implemented through deep learning(DL),which generally requires classification models to be highly accurate and robust.Aiming at the problems of inadequate visual management of emitter signals and low accuracy of signal recognition,this paper proposes a SEI method with signal feature knowledge graph(KG)and feature fusion.First,this method innovatively establishes signal feature database and uses KG to realize the visual representation of signal features.Secondly,feature fusion is carried out based on the constructed signal feature KG,which effectively improves the classification and recognition accuracy of specific emitters.The simulation results show that the proposed SEI method based on KG and feature fusion can visually manage the emitter signals better and improve the performance of SEI.