University of Tehran

Over the last decade, scholarly focus has intensified on optimizing and securing cyberphysical systems, particularly within the purview of transportation systems. This encompasses the refinement of vehicular equipment, constituting a subset of cyber-physical systems, with multifaceted goals encompassing heightened travel security, passenger comfort, and judicious resource utilization. Notably, the advent of Vehicular Ad-hoc Networks (VANETs) stands as a pivotal innovation aimed at augmenting travel safety and efficiency within intelligent transportation systems. These networks facilitate real-time vehicle-to-vehicle and vehicle-to-infrastructure communication, fostering environmental consciousness. However, safeguarding the optimal functioning of VANETs from malicious interventions, including the propagation of falsified messages, necessitates innovative approaches.

This study aims to introduce intelligent transportation systems and the challenges and limitations of vehicular networks. In addition, a novel method for the identification of malicious nodes within the VANET through post-processing of reported events is proposed. Drawing upon Bayesian theory, this methodology aligns with the inherent characteristics of vehicular networks. The proposed method aggregates latent information from received messages and subsequently computes the probability of node anomalies, thereby facilitating the classification of network nodes based on these probabilities. In terms of accuracy, our method achieved high accuracy with a limited corpus of received messages, rendering it particularly convenient in real-time automotive networks. Additionally, the approach exhibits the potential of the isolation of malicious nodes in future events, consequently, malicious nodes cannot reduce honest node reputation in the long term.

Keywords Intelligent Transportation Systems, Connected Vehicles, Intrusion Detection, Trust Management, Trust Model, Smart Cities, Vehicular Ad-hoc network (VANET), Cyber Security, Cognitive Networking.

My Proposal in Detail