Vehicle CAN Bus Message Abnormal Detection Based On AdaBoost Algorithm

With the continuous development and application of smart driving,smart transportation,cloud computing and 5G technologies,the automobile is not a simple travel tool,and it is gradually developing towards network and intelligence.Bring a more comfortable and convenient experience to people's lives.However,with the increasing number of external communication interfaces of automobiles,the interaction between the internal network of the automobile and the external network has become more and more frequent,which has led to more and more exposed loopholes in the Internet of Vehicles,which has brought many potential information security risks.Many car information security issues have also been continually leaked,and the network information security of the network has received great attention from all walks of life.CAN is the most commonly used bus for automobiles.It adopts multicast communication and has no corresponding security measures.The message data field is vulnerable to tampering and other attacks.For the problem that the data content of the CAN bus number message domain is abnormal,a car based on AdaBoost algorithm is proposed.CAN bus message anomaly detection method.The main work of this paper is as follows:(1)Analyzed the hidden dangers of CAN bus protocol and analyzed the reasons why CAN bus network is vulnerable to attack,summarized the attack modes and characteristics of common car network attacks.(2)For the problem of tampering attack in the data field of CAN bus message,the CART decision tree is used as the basic weak classifier,and the vehicle CAN bus packet anomaly detection model based on AdaBoost algorithm is proposed.Combined with the characteristics of the CAN bus message data field,64-bit data of the message data field is characterized by every 8 bits.The data pre-processing is used to process the packets with more serious or incorrect data,and the normal CAN bus data message and the abnormal CAN bus data message are distinguished according to the proposed abnormality detection method.(3)Using the real vehicle dataset,the algorithm is simulated and verified,and compared with the random forest model algorithm.Experiments show that the anomaly detection model proposed in this paper has a good detection effect on the data tampering attack of CAN bus network data.It is better than the random forest algorithm in the number of algorithm iterations and the number of classifiers.It is more suitable for the resource-restricted vehicle network environment.