Sybil attack detection in ultra-dense VANETs using verifiable delay functions

Abstract

Vehicular Ad Hoc Networks (VANETs) play a critical role in the future development of Intelligent Transportation Systems (ITS). These networks facilitate communication between vehicles and roadside infrastructure, establishing a dynamic network capable of sharing and processing traffic data. By harnessing this data, a comprehensive understanding of traffic conditions can be achieved, ultimately improving road safety and efficiency. VANETs have the potential to warn drivers about potential hazards, suggest optimal routes, and coordinate traffic signals. However, the current system design poses a vulnerability where a vehicle can acquire multiple identities, allowing it to launch a Sybil attack by impersonating multiple vehicles. In this attack, Sybil (or fake) vehicles generate and report false data, leading to fabricated congestion reports and corrupting traffic management data. To address this issue, this research proposes a novel Sybil attack detection scheme that leverages Verifiable Delay Functions (VDFs) and location data. The proposed scheme utilizes VDFs iteratively computed by vehicles throughout their journeys, forming a VDF chain where the included data is immutable. A vehicle obtains a signature on its recent VDF state from nearby Roadside Units (RSUs) and other vehicles and incorporates these signatures into its VDF chain. The inclusion of signatures in the VDF chains is time-bound and can’t be altered later. Essentially, the VDF chain serves as an immutable storage mechanism for each vehicle. Interactions between vehicles involve the exchange of signatures on VDF states, and these interactions, when compiled in a VDF chain, constitute a vehicle’s trajectory. By analyzing these trajectories, we can effectively detect Sybil trajectories. Unlike existing methods that solely rely on vehicle-to-RSU interactions, resulting in high false positive rates, our approach introduces vehicle-to-vehicle interactions using VDF chains, thereby increasing the detection rate. Extensive experiments and simulations are conducted to evaluate the proposed scheme’s performance in detection. The results demonstrate that our approach can accurately detect Sybil attacks while achieving low rates of false negatives and false positives when compared to existing models.