Cosine similarity based anomaly detection methodology for the CAN bus
作者：
Highlights：
• We proposed a cosine similarity based anomaly detection methodology for CAN bus.
• We evaluated our performance by considering the driving condition of the vehicle.
• The light-weight feature vector enables the method in minimizing detection time.
• The cosine similarity value gives clear boundary for detecting an abnormality.
• The abnormality could be identified with high accuracy and fast detection speed.
摘要
•We proposed a cosine similarity based anomaly detection methodology for CAN bus.•We evaluated our performance by considering the driving condition of the vehicle.•The light-weight feature vector enables the method in minimizing detection time.•The cosine similarity value gives clear boundary for detecting an abnormality.•The abnormality could be identified with high accuracy and fast detection speed.
论文关键词：In-vehicle network,Anomaly detection,Cosine similarity,Self-similarity
论文评审过程：Received 13 May 2019, Revised 29 June 2020, Accepted 27 September 2020, Available online 12 October 2020, Version of Record 5 November 2020.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.114066
