CABE-Trust: A Context-Aware Behavior Evaluation Model for Secure Communication in Autonomous UAV Networks

Abstract

Unmanned Aerial Vehicles (UAVs) are increasingly deployed in critical applications including environmental monitoring, military reconnaissance, and emergency response. These operations rely on autonomous coordination between UAVs through continuous inter-drone communication. Traditional cryptographic security mechanisms are insufficient in detecting internal threats from compromised, yet authenticated, UAVs. This paper introduces CABE-Trust (Context-Aware Behavior Evaluation for Trust), a decentralized trust framework that enables each UAV to assess the validity of received messages based on contextual semantics and behavioral history. CABE-Trust computes trust scores by evaluating three key dimensions: spatiotemporal consistency, semantic correctness, and historical reliability. Messages deemed inconsistent or suspicious reduce the sender’s trust level, triggering defensive actions such as message rejection or local broadcasting of alerts. The model is implemented in OMNeT++ and tested in multi-agent scenarios, including message injection, replay, and coordinated collusion. Experimental results demonstrate that CABE-Trust achieves a detection rate of 96.2% code, reduces false positives to 3.1% code, and adapts effectively in dynamic swarm environments, while introducing minimal computational and communication overhead. These findings establish CABE-Trust as a lightweight, scalable, and context-sensitive solution for securing UAV networks against advanced threats.