基于信号特征知识图谱与宽度学习架构的特定辐射源识别

Specific Emitter Identification Based on Signal Feature Knowledge Graph and Broad Learning Architecture

特定辐射源识别是一种有效的工业物联网数据流保护技术,现有的SEI方法可以通过机器学习来实现,而主流的SVM算法针对大规模训练样本会消耗大量的机器内存和运算时间。为了解决上述问题,提出了一种信号特征知识图谱与宽度学习架构的SEI方法KG-BLS算法,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。此外,该方法依靠特征映射节点和增强节点的非线性变换,通过单层前向传播网络来求解复杂的分类问题。仿真结果表明,所提出的KG-BLS算法在识别性能和计算开销方面具有巨大的优势。

Specific emitter identification(SEI)is an effective data stream protection method for the Industrial Internet of Things.Existing SEI methods can be implemented through machine learning(ML),while mainstream SVM algorithms will consume a lot of machine memory and computing time for large-scale training samples.In order to solve the above problem,this paper proposes a SEI method based on signal feature knowledge graph(KG)and broad learning architecture,i.e.,KG-BLS algorithm.This method innovatively establishes a signal feature database,and uses the KG to realize visual representation of signal features.In addition,the method relies on the nonlinear transformation of feature mapping nodes and enhancement nodes,and solves complex classification problems through a single-layer forward propagation network.The simulation results show that the proposed KG-BLS algorithm has great advantages in recognition performance and computational cost.