Cybersecurity for Connected Vehicle Networks: Leveraging Sampled Network Traffic Beyond the CAN Protocol

As connected vehicles evolve, ensuring robust cybersecurity becomes critical. Traditional anomaly detection methods in automotive networks have primarily focused on the CAN protocol, which is limited in scope and does not account for the complexities of modern, high-bandwidth vehicle networks and sophisticated cyberattacks. This paper introduces a novel approach that analyzes small samples of network traffic packets rather than CAN bus data. The method was tested in collaboration with a leading automotive manufacturer, demonstrating its capability to detect a wide range of attacks. The results show that this approach offers significant advantages in scalability, privacy, and real-time cybersecurity for connected vehicle systems.