Sybil Attack Detection Based on Signal Clustering in Vehicular Networks

With the growing adoption of vehicular networks, ensuring the security of these networks is becoming increasingly crucial. However, the broadcast nature of communication in these networks creates numerous privacy and security concerns. In particular, the Sybil attack, where attackers can use multiple identities to disseminate false messages, cause service delays, or gain control of the network, poses a significant threat. To combat this attack, we propose a novel approach utilizing the channel state information (CSI) of vehicles. Our approach leverages the distinct spatio-temporal variations of CSI samples obtained in vehicular communication signals to detect these attacks. We conduct extensive real-world experiments using vehicle-to-everything (V2X) data, gathered from dedicated short-range communications (DSRC) in vehicular networks. Our results demonstrate a high detection rate of over 98% in the real-world experiments, showcasing the practicality and effectiveness of our method in realistic vehicular scenarios. Furthermore, we rigorously test our approach through advanced ray-tracing simulations in urban environments, which demonstrates high efficacy even in complex scenarios involving various vehicles. This makes our approach a valuable, hardware-independent solution for the V2X technologies at major intersections.