CoDetect: cooperative anomaly detection with privacy protection towards UAV swarm

In this study, we introduce a novel framework for cooperative anomaly detection in UAV swarms. The scheme integrates an anomaly detection model, consensus algorithm, and lightweight communication authentication algorithm. Tailored to address external eavesdroppers and malicious Byzantine nodes, it effectively manages and mitigates Byzantine behavior while safeguarding internal communication. Simultaneously, the framework incorporates a lightweight authentication scheme designed to verify node legitimacy and enhance swarm scalability. Compared with existing schemes, it demonstrates competitiveness in communication, computing costs, and consensus algorithm efficiency. Consequently, we assert that the proposed framework is effective and feasible for swarm anomaly detection. However, in more intricate scenarios, there is room for further refinement. Therefore, delineating attack dictionaries and formulating defense strategies emerge as noteworthy future directions for this work.