CATS: A framework for Cooperative Autonomy Trust & Security

TL;DR

CATS addresses the security challenges of cooperative perception in V2X by integrating long-term reputation with in-situ majority validation, mediated by a centralized Security Authority. The system uses a dual-rate voting mechanism with rate limits and a three-state trust model to quickly exclude misbehaving vehicles while preserving availability and privacy through pseudonyms. City-scale experiments and a formal risk analysis show that CATS can reduce bad messages by roughly two orders of magnitude and ban misbehaving vehicles rapidly (around tens of seconds) with minimal impact on legitimate data, while certificate verification remains fast. These results indicate that a hybrid, privacy-preserving trust framework can be realistically deployed in near-future autonomous-vehicle networks, with clearly defined parameters and potential avenues for decentralization and enhancement.

Abstract

With cooperative perception, autonomous vehicles can wirelessly share sensor data and representations to overcome sensor occlusions, improving situational awareness. Securing such data exchanges is crucial for connected autonomous vehicles. Existing, automated reputation-based approaches often suffer from a delay between detection and exclusion of misbehaving vehicles, while majority-based approaches have communication overheads that limits scalability. In this paper, we introduce CATS, a novel automated system that blends together the best traits of reputation-based and majority-based detection mechanisms to secure vehicle-to-everything (V2X) communications for cooperative perception, while preserving the privacy of cooperating vehicles. Our evaluation with city-scale simulations on realistic traffic data shows CATS's effectiveness in rapidly identifying and isolating misbehaving vehicles, with a low false negative rate and overheads, proving its suitability for real world deployments.

Figures (12)

• Figure 1: Phantom red light violator: a data injection attack.
• Figure 2: CATS system workflow.
• Figure 3: CATS architecture at a component level, showing how individual pieces fits together.
• Figure 4: State diagram for CATS (transition descriptions in middle). Trust state determination (right) is run after every transition.
• Figure 5: Comparative analysis of false negative and false positive message percentages for misbehavior detection methods (CATS, VANET Reputation with $\psi_{nf} = 2$ and $\psi_{nf} = 100$, and TruPercept) in Berlin and Boston traffic simulation runs.