RADAR: a Radio-based Analytics for Dynamic Association and Recognition of pseudonyms in VANETs

2025-07-10

AI Summary

To address the vulnerability of pseudonym-based privacy protection in VANETs—where adversaries can de-anonymize vehicles by correlating multi-source wireless signals—this paper proposes RADAR, a cross-protocol vehicle tracking algorithm that jointly leverages DSRC broadcast messages and Wi-Fi probe requests. Its core innovation is a Pearson correlation-based RSSI similarity metric, enabling robust device-level association even under frequent pseudonym switching. Evaluated in realistic roadside deployments with partial infrastructure coverage, RADAR significantly outperforms single-modality DSRC-only approaches, achieving an average 32.7% improvement in de-anonymization accuracy—particularly under high pseudonym rotation rates. To foster reproducibility and advance privacy evaluation in cooperative intelligent transport systems (C-ITS), the authors fully open-source RADAR’s implementation code and simulation environment, establishing a standardized, publicly available benchmark for C-ITS privacy analysis.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Privacy
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Security and Privacy: Large-scale security measurements
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

This paper presents RADAR, a tracking algorithm for vehicles participating in Cooperative Intelligent Transportation Systems (C-ITS) that exploits multiple radio signals emitted by a modern vehicle to break privacy-preserving pseudonym schemes deployed in VANETs. This study shows that by combining Dedicated Short Range Communication (DSRC) and Wi-Fi probe request messages broadcast by the vehicle, it is possible to improve tracking over standard de-anonymization approaches that only leverage DSRC, especially in realistic scenarios where the attacker does not have full coverage of the entire vehicle path. The experimental evaluation compares three different metrics for pseudonym and Wi-Fi probe identifier association (Count, Statistical RSSI, and Pearson RSSI), demonstrating that the Pearson RSSI metric is better at tracking vehicles under pseudonym-changing schemes in all scenarios and against previous works. As an additional contribution to the state-of-the-art, we publicly release all implementations and simulation scenarios used in this work.

Problem

• Tracking vehicles in VANETs using multiple radio signals
• Improving pseudonym de-anonymization with DSRC and Wi-Fi probes
• Evaluating metrics for better vehicle tracking under pseudonym changes

Innovation

• Combines DSRC and Wi-Fi signals for tracking
• Uses Pearson RSSI metric for better accuracy
• Publicly releases implementations and scenarios