LA-DETECTS: Local and Adaptive Data-Centric Misbehavior Detection Framework for Vehicular Technology Security

Abstract

Vehicular Ad Hoc Networks (VANET) represent an immense technological advancement enhancing connectivity among Vehicular Technology including vehicles and roadside infrastructure to ensure road safety and improve forthcoming transportation services. The effectiveness of safety applications depends on the reliability and consistency of periodically broadcasted real-time environmental and vehicle state information. However, insider threats arise when nodes with valid access credentials disseminate maliciously incorrect information. Existing misbehavior detection solutions are often static and lack the adaptability required for the dynamic nature of vehicular networks, leaving a gap in addressing sophisticated attacks such as Denial of Service (DoS), data replay, and Sybil attacks. To fill this gap, we propose a context-aware, data-driven misbehavior detection framework that allows each vehicle to perform plausibility and consistency checks on received messages. The Adaptive Misbehavior Detection Framework addresses critical security challenges within localized vehicles by incorporating dynamically computed parameters and confidence intervals to assess message integrity. To determine the presence of misbehavior, a weighted average approach effectively reduces the possibility of false positives. Simulation results demonstrate that our proposed mechanism significantly enhances detection performance against key misbehavior types, including false information dissemination, DoS, disruptive, and variants of Sybil attacks variants, outperforming existing benchmarks with the VeReMi extension dataset.

Keywords

• Vehicular technology
• security
• attacks
• misbehavior detection framework
• kalman filter
• plausibility and consistency checks