Research on Vehicle Network Intrusion Detection System Based on K-means Clustering Algorithm

Abstract

As the most widely used network in the vehicles, the Controller Area Network (CAN) is vulnerable to cyber-attacks due to the limitations of the bandwidth and transmission rate, lack of identity authentication, data encryption and other primary protection functions. Based on the known defects of intrusion detection systems, this paper shows a time voltage hybrid detection scheme that only relies on the physical information on the CAN bus when the data information has become unreliable. A lightweight K-means algorithm is used to train and verify the data. The test results show that the intrusion detection scheme in this paper can effectively detect the number of Electronic Control Units (ECUs) on the CAN network and play a role in checking the admission of the external ECUs connected to the bus.