Lightweight Diagnostic-based Secure Framework for Electronic Control Units in Vehicles

Cyber-physical security has become a central and primary concern in the automotive field due to its rapid development. Thus, the necessity to introduce intrusion detection system (IDS) has raised recently to counter the possible malicious attacks. However, the location of the IDS in the vehicle network is considered a problem due to several limitations. We introduce a novel design for a lightweight intrusion detection framework. Gradient boosting (XGBoost) technique is used in the proposed framework to detect the anomalies in diagnostics data. Experimental results show that the model is capable to detect anomalies with reasonable false acceptance rate (FA) and high accuracy. Also, the results highlight the superiority of the model against the others. Additionally, to show that the model is robust and stable to be used for different datasets, two datasets collected from real vehicles are used. Moreover, an attack generation algorithm is proposed to introduce the anomalies in the datasets.