Enhancing CAN-FD Security with an Ensemble Learning Approach for Intrusion Detection

As the integration of advanced technologies in vehicular systems continues to rise, the risk of cyber threats to in-vehicle networks is also becoming a critical concern. Intrusion detection systems (IDS) play a crucial role in safeguarding these networks against unauthorized access and malicious activities. Traditional IDS approaches often face challenges in effectively detecting cyber threats due to the technological advancements in automotive environments. This study proposes an ensemble learning-based intrusion detection model tailored for CAN-FD (Controller Area Network - Flexible Datarate) networks. The model combines the strength and computational simplicity of multiple classification algorithms, including Random Forest (RF), Decision Trees (DT), and k-Nearest Neighbors (KNN), to create a robust and adaptive detection system. Through comprehensive experimentation and evaluation with the benchmark CAN-FD intrusion dataset, the proposed ensemble model demonstrates the reliability of the model compared to individual classifiers. This research contributes to in-vehicle security by providing an effective and scalable approach to intrusion detection in CAN-FD networks.