无线电工程2025，Vol.55Issue(9)：1861-1868,8.DOI:10.3969/j.issn.1003-3106.2025.09.013
无人机辅助联邦学习安全传输能效优化
Secure and Energy-efficient Transmission Based on UAV-assisted Federated Learning
摘要
Abstract
Federated learning is a privacy-preserving distributed machine learning paradigm that reduces communication overhead by transmitting model parameters instead of raw data.However,the practical deployment of federated learning still faces two major challenges.Firstly,limited ground communication infrastructure restricts efficient collaboration among devices;secondly,the communication process,especially in open environments,is vulnerable to eavesdropping,increasing risks of privacy leakage.To address these issues,UAV are utilized as aerial mobile access platforms to provide temporary communication support for edge devices and effectively mitigate the problem of incomplete communication coverage.Furthermore,a security-aware joint resource optimization scheme is proposed,which maximizes energy efficiency while ensuring secure transmission by jointly optimizing the CPU frequency,transmit power,and local model accuracy of user devices.The formulated non-convex optimization problem is decomposed into three subproblems using an alternating optimization strategy,which are then solved via fractional programming and Successive Convex Approximation(SCA)methods to obtain a near-optimal solution.Simulation results demonstrate that the proposed approach can significantly enhance system security performance while effectively reducing overall energy consumption compared to existing baseline schemes.关键词
无人机/联邦学习/安全传输/资源优化Key words
UAV/federated learning/secure transmission/resource optimization分类
信息技术与安全科学引用本文复制引用
孟涛,窦艳菲,孙强,陈晓敏..无人机辅助联邦学习安全传输能效优化[J].无线电工程,2025,55(9):1861-1868,8.基金项目
国家自然科学基金(62401297)National Natural Science Foundation of China(62401297) （62401297）
