Research on Anomaly Detection of Vehicle-mounted Controller Area Network

In recent years, with the rapid development of technologies such as the Internet, big data, and artificial intelligence, intelligent connected vehicles have begun to appear in human life, and the security research of vehicle networks has gradually become a hot topic. The paper analyzes the characteristics and security issues of the vehicle controller area network, and proposes a method that combines the chi-square test and the Gentle Adaboost algorithm for vehicle controller area network intrusion detection. First, the chi-square test is used to select the features of the real data, and then the filtered data is imported into the integrated learning algorithm Gentle Adaboost for learning and detection. Experimental results show that our method performs better than several other classical algorithms.