Anomaly Detection in Cooperative Vehicle Perception Systems under Imperfect Communication

2025-01-28

AI Summary

To address the insufficient robustness of anomaly detection in multi-vehicle collaborative perception under communication constraints (e.g., low bandwidth and high disconnection rates), this paper proposes CPAD—the first communication-interruption-resilient collaborative perception anomaly detection framework. Methodologically, CPAD integrates vehicle dynamics-based rule modeling, trajectory-driven multi-agent anomaly classification, and a robust consensus mechanism. We further introduce the first large-scale, open-source multi-vehicle anomaly detection benchmark, comprising 90,000 trajectories across 15,000 realistic scenarios. Experiments demonstrate that CPAD significantly outperforms state-of-the-art methods in both F1-score and AUC, maintaining stable performance even under 40% connection interruption rate. Our core contributions are: (i) the first collaborative anomaly detection framework explicitly designed for communication-vulnerable environments; (ii) the first publicly available, large-scale multi-vehicle anomaly detection benchmark; and (iii) empirical validation that rule-guided modeling and consensus mechanisms are critical for enhancing system robustness against communication failures.

Technology Category

• Computer Vision: Multi-modal Vision
• Multiagent Systems: Agent Communication
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Security and Privacy: Large-scale security measurements
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

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.

Problem

• Abnormal Behavior Detection
• Incomplete Information Transmission
• Autonomous Vehicle Safety

Innovation

• Intelligent Transportation Systems
• Anomaly Detection
• Vehicular Communication