车联网环境下基于CNN-LSTM的行驶信息欺骗攻击检测  

Driving information spoofing attack detection based on CNN-LSTM in internet of vehicles

作　　者：梁乐威 陈宇峰 向郑涛[1] 游康祥 周旭 LIANG Lewei;CHEN Yufeng;XIANG Zhengtao;YOU Kangxiang;ZHOU Xu(School of Electrical and Information Engineering,Hubei University of Automotive Technology,Shiyan 442002,China;School of Automotive Engineers,Hubei University of Automotive Technology,Shiyan 442002,China)

机构地区：[1]湖北汽车工业学院电气与信息工程学院,湖北十堰442002 [2]湖北汽车工业学院汽车工程师学院,湖北十堰442002

出　　处：《江苏理工学院学报》2023年第6期31-39,共9页Journal of Jiangsu University of Technology

基　　金：教育部中国高校产学研创新基金项目“基于边缘计算的车路协同技术研究”(2021FNA04017)。

摘　　要：当联网车辆遭受网络攻击时,会向外广播虚假行驶信息,从而误导周边车辆,极易引发交通事故。针对这一问题,文章提出了一种基于一维卷积神经网络(CNN)和长短期记忆神经网络(LSTM)的组合深度学习模型,通过提取车辆行驶信息的有效特征对模型进行训练,并对车联网环境下的速度与位置等行驶信息的欺骗攻击进行检测。在VeReMi Extension数据集上验证了模型的有效性,实验结果表明:所提方案对比单一的CNN、LSTM模型在召回率指标上分别提升了8.7%和6.1%,在F1分值上分别提升了4.6%和4.1%。When the connected vehicles are subjected to network attacks,it will broadcast false driving information to mislead surrounding vehicles,which is very easy to cause traffic accidents.In order to solve this problem,a combined deep learning model based on a one-dimensional convolutional neural network(CNN)and a long short-term memory neural network(LSTM)is proposed.The model can be trained by extracting effective features of vehicle driving information,and the spoofing attacks of driving information such as speed and location in the networked vehicle environment are also detected.The effectiveness of the model has been verified on the VeReMi Extension dataset,the experimental results show that the proposed scheme improves 8.7%and 6.1%,respectively in recall rate,as well as increases 4.6%and 4.1%,respectively in F1 score compared to single CNN and single LSTM models.

关 键 词：车载自组织网络 攻击检测 深度学习 欺骗攻击 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]