Anomaly Detection in Cooperative Vehicle Perception Systems under Imperfect Communication

TL;DR

This paper introduces CPAD, a robust cooperative perception framework for anomaly detection in autonomous vehicles that remains effective under imperfect communication, supported by a large, newly created benchmark dataset.

Contribution

The paper presents a novel anomaly detection framework for cooperative vehicle perception that is resilient to communication failures and introduces a large-scale benchmark dataset for multi-agent vehicle trajectories.

Findings

Outperforms standard anomaly detection methods in F1-score and AUC

Demonstrates robustness to communication interruptions

Provides a new large-scale vehicle trajectory dataset

Abstract

Anomaly detection is a critical requirement for ensuring safety in autonomous driving. In this work, we leverage Cooperative Perception to share information across nearby vehicles, enabling more accurate identification and consensus of anomalous behaviors in complex traffic scenarios. To account for the real-world challenge of imperfect communication, we propose a cooperative-perception-based anomaly detection framework (CPAD), which is a robust architecture that remains effective under communication interruptions, thereby facilitating reliable performance even in low-bandwidth settings. Since no multi-agent anomaly detection dataset exists for vehicle trajectories, we introduce 15,000 different scenarios with a 90,000 trajectories benchmark dataset generated through rule-based vehicle dynamics analysis. Empirical results demonstrate that our approach outperforms standard anomaly classification methods in F1-score, AUC and showcase strong robustness to agent connection interruptions.