摘要
光网络由于其结构的脆弱性,容易受到旨在中断通信服务的信号干扰攻击。基于此,提出了一种基于机器学习的攻击检测、识别与恢复框架。在攻击检测与识别方面,评估了BiLSTM、1DCNN和7种常规机器学习分类器(ANN、DT、KNN、LDA、NB、RF和SVM)在检测攻击是否存在,以及识别受到的不同类型的干扰攻击上的性能。在攻击恢复方面,提出了基于BiLSTM-BiGRU的干扰攻击恢复模型,分别用来恢复轻度带内、强度带内、轻度带外和强度带外干扰攻击。数值仿真结果表明,所提模型表现出优异的性能,检测与识别准确率高达99.20%,针对4种攻击的恢复率分别为95.05%、97.03%、94.06%和61.88%。
Optical networks are vulnerable to signal jamming attacks aimed at disrupting communication services due to their structural fragility.Based on this,a machine learning-based jamming attacks detection,identification and restoration framework was proposed.In terms of attacks detection and identification,the performances of BiLSTM,1DCNN,and seven conventional machine learning classifiers(ANN,DT,KNN,LDA,NB,RF,and SVM)were evaluated in detecting the presence of attacks,and identifying different types of jamming attacks.In terms of attacks restoration,a BiLSTM-BiGRU-based jamming attacks restoration model was proposed to restore light-in-band,strong-in-band,light-out-of-band,and strong-out-of-band jamming attacks.Numerical simulation results reveal that the proposed model demonstrates excellent performance with a detection and identification accuracy of 99.20%,with attack restoration ratios of 95.05%,97.03%,94.06%,and 61.88%,respectively.
作者
巩小雪
庞嘉豪
张琦涵
徐长乐
秦文帅
郭磊
GONG Xiaoxue;PANG Jiahao;ZHANG Qihan;XU Changle;QIN Wenshuai;GUO Lei(School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China;Institute of Intelligent Communication and Network Security,Chongqing University of Posts and Telecommunications,Chongqing 400065,China;School of Computer Science and Engineering,Northeastern University,Shenyang 110819,China)
出处
《通信学报》
EI
CSCD
北大核心
2023年第7期159-170,共12页
Journal on Communications
基金
国家自然科学基金资助项目(No.62075024,No.62025105,No.62201105,No.62205043,No.62221005)
重庆市自然科学基金资助项目(No.CSTB2022NSCQ-MSX1334,No.cstc2021jcyj-msxmX0404)
重庆市教委创新研究群体基金资助项目(No.CXQT21019)。
关键词
机器学习
攻击检测与识别
攻击恢复
光网络安全
machine learning
attack detection and identification
attack recovery
optical network security
