An intrusion detection method for the in-vehicle network

In view of the defects of the vehicle controller area network (CAN) bus protocol and the large number of attack interfaces exposed by the vehicle, the vehicle network is vulnerable to network attacks, which leads to vehicle information security and driving safety issues. And current intrusion detection algorithms are difficult to achieve realtime online intrusion detection in the in-vehicle network. This paper proposes an in-vehicle CAN network intrusion detection method based on message ID and message cycle. First, this method performs whitelist detection of message ID, and then performs dual-threshold intrusion detection based on message cycle. The experimental results show that the detection accuracy of this method against in-vehicle CAN network message injection attacks is 94.13%, the precision rate is 92.46%, and the false alarm rate is 4.01 %.