一种基于支持向量机的车载网络异常检测方法

随着人工智能、5G、激光雷达和各类传感器等技术的不断发展与应用,无人驾驶、车联网等应运而生,汽车朝着智能化和网联化不断发展,为人们带来舒适、安全的驾驶体验。同时,网联化也打破了汽车现有的闭环状态,为车载电子系统带来了潜在的信息安全问题。为此,文章提出了基于支持向量机的车载网络入侵检测算法。通过对报文的DATA域的分析,挖掘报文的各字节特点,综合各字节和字节数据的信息熵,构成分类训练样本,训练支持向量模型,以此检测数据的可能异常。通过真实车辆数据实验分析,对模拟攻击的异常检测具有较高的检测率。

An Anomaly Detection Method Based on Support Vector Machine for In-vehicle Network

With the development and application of artificial intelligence,5G,lidar and every kinds of sensors and other technologies,driveless and vehicle internet come alive.Vehicles is becoming more and more intelligent and networking,bringing us comfort and safe experience of driving.At the same time,networking also breaks the existing close-loop state of the vehicle,which brings the potential problems of information safety.Therefore,our paper proposes a vehicle network intrusion detection algorithm based on support vector machine.Through analyzing the DATA domain of the message,digging out the characters of each byte of the message,synthesizing the information entropy of each byte and byte data,forming the classified training sample,training support vector machine model,we can detect the possible abnormalities of data.By analyzing the real vehicle data experiment,the anomaly detection of simulated attack has a high detection rate.