Federated Learning for Trust Enhancement in UAV-Enabled IoT Networks: A Unified Approach

Abstract

This study presents a federated learning (FL) framework tailored for uncrewed-aerial-vehicle (UAV)-enabled Internet of Things (IoT) networks, addressing challenges in efficiency, robustness, and scalability. The proposed system improves model learning with a 14.9 percentage point increase in accuracy (75.5%–90.4%) and a 69.2% reduction in loss over ten training epochs. It demonstrates resilience, limiting accuracy reduction to 7% under simulated attacks, and scalability with a linear increase in processing times as network size grows. High anomaly detection rates (92%) further enhance network security and reliability. These results validate the framework’s effectiveness in UAV networks and highlight its broader potential for IoT applications. Future work will explore further enhancements and diverse applications.