Machine learning-based detection and mitigation of cyberattacks in adaptive cruise control systems

The growing reliance on Vehicle-to-Vehicle (V2V) communication has heightened the vulnerability of Adaptive Cruise Control (ACC) systems to cybersecurity threats, such as manipulation or forgery of V2V messages. This paper investigates the impact of three types of false information injection (FII) on vehicle collision risk and driving efficiency. To address these vulnerabilities, we develop a novel machine learning-based onboard model, ACC anomaly Detection and Mitigation (ACCDM), designed to strengthen ACC resilience against such cyberattacks. ACCDM continuously monitors vehicle parameters under benign conditions, detecting deviations that indicate potential threats and deploying real-time mitigations to maintain safety and efficiency. Simulations across continuous and clustered attack scenarios validate ACCDM's accuracy in detecting cybersecurity threats, preserving safe following distances, and mitigating the negative impacts of cyberattacks on ACC systems.