UASDefMeta: A Meta-Learning-Based Defense Approach for Detecting Unmanned Aerial Systems Eavesdropping in Mission-Critical Applications

Unmanned aerial systems (UASs) play a vital role in various mission-critical applications, such as military missions, surveillance, and emergency responses. Ensuring secure communication channels for UASs is crucial to protecting sensitive information and maintaining operational integrity. Eavesdropping attacks pose a significant threat to data confidentiality and system security in these applications. To address these challenges, we propose an environment in Python that emulates eavesdropping attacks on UASs. Subsequently, we introduce UASDefMeta, a meta-learning-based approach that combines model-agnostic meta-learning and proximal policy optimization for detecting these attacks. Experiments validate UASDefMeta’s effectiveness, demonstrating its adaptability and improved performance over other selected models.