Research Article
Research on Resource Management Algorithms Based on Federated Learning in UAV-Assisted MEC
@INPROCEEDINGS{10.1007/978-3-032-33662-0_28, author={Pu Zhao and Yang Lou and Chenyu Meng and Li Wang}, title={Research on Resource Management Algorithms Based on Federated Learning in UAV-Assisted MEC}, proceedings={Communications and Networking. 20th EAI International Conference, ChinaCom 2025, Xi’an, China, November 8--9, 2025, Proceedings}, proceedings_a={CHINACOM}, year={2026}, month={10}, keywords={Federated learning Deep reinforcement learning MEC Unmanned aerial vehicle}, doi={10.1007/978-3-032-33662-0_28} }- Pu Zhao
Yang Lou
Chenyu Meng
Li Wang
Year: 2026
Research on Resource Management Algorithms Based on Federated Learning in UAV-Assisted MEC
CHINACOM
Springer
DOI: 10.1007/978-3-032-33662-0_28
Abstract
As Mobile Edge Computing (MEC) becomes increasingly important in edge data processing, Unmanned Aerial Vehicles (UAVs) have emerged as key enablers for MEC due to their flexible deployment and broad coverage. However, UAVs are constrained in computing and communication resources, and the presence of imperfect Channel State Information (CSI) further complicates resource management. Federated Learning (FL) offers a promising solution by enabling efficient coordination between UAVs and base stations. This paper investigates FL-based dynamic resource management in UAV-assisted MEC under imperfect CSI. We model the joint optimization of UAV selection, transmission power, and CPU frequency as a Markov Decision Process (MDP), and apply the Std Q-target (SQT) framework for online policy learning. Our proposed FL-SQT algorithm improves the success rate by 1.98% and the Accuracy-to-Energy (AE) gain by 29.2% compared to traditional methods.