| In recent years,the intellectualization and interconnection of automobile obtained a great development,a large number of intelligent electronic control units were integrated in automobile which made vehicles can automatic drive.The electronic control units with wireless communication function in automobile make vehicle become a node of Internet,the automobile became a “computer” with four wheels.Meanwhile,the vulnerability of network security in automobile can be used by hacker to intrude in-vehicle network which can cause traffic accidents and injuries.It is urgent to research security of in-vehicle network.The research of security strategy for in-vehicle network was needed to be done according to the feature of in-vehicle network to keep safely driving.The status of in-vehicle network security in the world are researched in this paper,the problems of real time and incomplete detection in existing security strategies are pointed out,according to the characteristics of vehicle network,two intrusion detection methods based on detection of key feature and detection of rule were designed.The hardware platform of the intrusion detection system was built,and the intrusion detection methods proposed in this paper were implemented in the hardware platform.Finally,the system is tested and verified in test environment of hardware-in-loop.The main contents of this paper are as follows:1.The security of network was be researched from structure and communication protocol aspects.The security flaws of network were pointed out in this research.The intrusion detection methods for CAN were proposed according to the research of technology of intrusion detection system and introduce some relevant theories.2.An intrusion detection method based on key features was proposed.The attack methods of CAN network were analyzed according to security flaws of CAN network in vehicle.The features which were influenced by attack were extracted as key features.The normal modes of these key features were built and an intrusion detection method based on key features detection was design according to these normal modes,a simulation model was built to verify the method.3.An intrusion detection method based on rule was proposed.Collecting the CAN messages from real vehicle.the method of association rules mining was adopted to mine rules between signals and design the signal rule dataset according to relationship between the signals.An intrusion detection method based on signal rule detection was design according to signal rule dataset and a simulation analysis was carried out to verify the method.4.The hardware-in-loop test environment was design based on real vehicle network structures and design a network attack tool.An intrusion detection system hardware was design.two methods of intrusion detection proposed in this paper were implemented in the intrusion detection system hardware.The intrusion detection system designed in this paper was test in this environment.The experimental data shows that the detection rate of intrusion detection system designed in this paper was 96.24% and the detection response time within 9 ms,which can effectively increase the security of in-vehicle network. |
