Network Intrusion Detection in Autonomous Vehicles

Autonomous vehicles promise safer and more efficient transportation, yet they face cybersecurity risks due to their reliance on interconnected systems. This paper emphasizes and sizes the critical role of intrusion detection systems (IDS) in safeguarding these vehicles. Various Intrusion Detection System (IDS) methods are explored and designed to meet the specific security needs of Controller Area Networks (CAN) in autonomous cars, emphasizing machine learning techniques. Our paper provides significant performance improvements in IDS through feature selection and hyperparameter optimization. Experimental results show a boost in the F1 score from 98.861 % to 99.57%, with a 70.2% reduction in execution time. Additionally, the IDS achieves 99.57% accuracy in binary classification. These findings highlight the importance of optimizing IDS for robust security in autonomous vehicle environments.