Security Research Of CAN Bus For Intelligent And Connected Vehicles

With the rapid popularization of communication network,various emerging technologies are gradually mounted in cars,which makes the intelligent and connected vehicles(ICVs)develop rapidly and gradually integrate into people’s daily life.Although a variety of emerging technologies provide users with comfortable service and good driving experience,at the same time,they also bring a lot of available communication interfaces for cars,which makes cars vulnerable to hackers.It is found that although there are many ways for malicious attackers to invade the car,the last step of the attack must be through the in-vehicle network,namely controller area network(CAN)bus.Due to the weakness of access control,lack of encryption and message authentication,CAN bus is vulnerable to malicious intrusion.Therefore,the research of vehicle attack experiment and related defense technology based on CAN bus vulnerability is of great significance to protect vehicle security.In addition to the in-vehicle network security problems,there are other potential safety problems need to be solved.For example,driving fingerprint authentication.In the existing identity authentication methods,static password,smart card and other authentication methods have the risk of leakage,while biometric methods such as face recognition and iris recognition cannot monitor the driver’s identity in real time.Therefore,it is of great significance to design a real-time vehicle driver fingerprint authentication system to prevent unauthorized drivers from driving illegally.In addition,how to reduce the occurrence of traffic accidents has been a hot research topic in the process of automobile development.It is found that the driver is the decisive factor in traffic accidents,and the decision-making and behavior of the driver,such as the distance between the driver and the vehicle,whether to change lanes and overtake,depend on the driver’s driving habits.Drivers with poor driving behavior and unskilled driving skills may cause traffic accidents by misoperation in a specific situation.Therefore,it is a good research direction to find a method that can evaluate the driver’s driving behavior habits and let the driver actively improve the driving behavior.In view of the above problems,this dissertation focuses on the following four aspects.1.Experimental research on vehicle attack based on CAN bus vulnerability.Aiming at the weakness of CAN bus,such as weak access control,no effective authentication mechanism,weak encryption protection and so on,this dissertation carries out the experimental research on the attack of CAN bus for Luxgen automobile,and realizes the remote control of automobile dashboard,door lock,lamp and other components.The experimental results show that there are many exploitable loopholes in the car,which seriously threaten the life safety of drivers and passengers.2.Design of CAN bus intrusion detection system based on vehicle voltage signal.Aiming at the weakness of CAN bus security defense,a CAN bus intrusion detection system based on ECU voltage signal characteristics is designed.The system does not occupy the bandwidth and computing resources of CAN bus,and is not interfered by vehicle status,the number of attacking ECU and attack data transmission frequency.Once the external intrusion devices are connected to CAN bus to send attack messages,the system can achieve high accuracy detection.3.Driving fingerprint authentication method based on side channel analysis.In view of the hidden danger in special vehicles such as cash vehicles,trucks and buses,this dissertation designs a real-time anti hijacking driver authentication method based on the extraction of driving behavior characteristics,combining with the side channel analysis technology.The experimental results show that the method can achieve high accuracy for the recognition of authorized drivers and unauthorized drivers.4.Driving behavior evaluation method based on transfer learning.In order to improve the driver’s driving level and reduce the occurrence of traffic accidents,this dissertation proposes a new driving behavior evaluation method based on transfer learning.This scheme can transfer the knowledge from the old domain to the new domain,and realize the grading and evaluation of driver’s driving behavior.