Decentralized Aggregation for Energy-Efficient Federated Learning in mmWave Aerial-Terrestrial Integrated Networks

被引:0
|
作者
Saif, Mohammed [1 ]
Hassan, Md. Zoheb [2 ]
Hossain, Md. Jahangir [3 ]
机构
[1] University of Toronto, Department of Electrical and Computer Engineering, Toronto,ON,M5S 1A1, Canada
[2] Université Laval, Electrical and Computer Engineering Department, Québec,QC,G1V 0A6, Canada
[3] The University of British Columbia, School of Engineering, Kelowna,BC,V1V 1V7, Canada
来源
IEEE Transactions on Machine Learning in Communications and Networking | 2024年 / 2卷
关键词
Energy utilization;
D O I
10.1109/TMLCN.2024.3410211
中图分类号
学科分类号
摘要
It is anticipated that aerial-terrestrial integrated networks incorporating unmanned aerial vehicles (UAVs) mounted relays will offer improved coverage and connectivity in the beyond 5G era. Meanwhile, federated learning (FL) is a promising distributed machine learning technique for building inference models over wireless networks due to its ability to maintain user privacy and reduce communication overhead. However, off-the-shelf FL models aggregate global parameters at a central parameter server (CPS), increasing energy consumption and latency, as well as inefficiently utilizing radio resource blocks (RRBs) for distributed user devices (UDs). This paper presents a resource-efficient and decentralized FL framework called FedMoD (federated learning with model dissemination), for millimeter-wave (mmWave) aerial-terrestrial integrated networks with the following two unique characteristics. Firstly, FedMoD incorporates a novel decentralized model dissemination scheme that uses UAVs as local model aggregators through UAV-to-UAV and device-to-device (D2D) communications. As a result, FedMoD 1) increases the number of participant UDs in developing the FL model; and 2) achieves global model aggregation without involving CPS. Secondly, FedMoD reduces FL's energy consumption using radio resource management (RRM) under the constraints of over-the-air learning latency. To achieve this, by leveraging graph theory, FedMoD optimizes the scheduling of line-of-sight (LOS) UDs to suitable UAVs and RRBs over mmWave links and non-LOS UDs to available LOS UDs via overlay D2D communications. Extensive simulations reveal that FedMoD, despite being decentralized, offers the same convergence performance to the conventional centralized FL frameworks. © 2023 CCBY.
引用
收藏
页码:1283 / 1304
相关论文
共 50 条
  • [21] Computation Offloading and Resource Allocation in MEC-Enabled Integrated Aerial-Terrestrial Vehicular Networks: A Reinforcement Learning Approach
    Waqar, Noor
    Hassan, Syed Ali
    Mahmood, Aamir
    Dev, Kapal
    Dinh-Thuan Do
    Gidlund, Mikael
    IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 2022, 23 (11) : 21478 - 21491
  • [22] Federated Learning Model Aggregation in Heterogenous Aerial and Space Networks
    Dong, Fan
    Abbasi, Ali
    Leung, Henry
    Wang, Xin
    Zhou, Jiayu
    Drew, Steve
    2024 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS WORKSHOPS, ICC WORKSHOPS 2024, 2024, : 1262 - 1267
  • [23] Optimizing federated learning with weighted aggregation in aerial and space networks
    Dong, Fan
    Leung, Henry
    Drew, Steve
    JOURNAL OF NETWORK AND COMPUTER APPLICATIONS, 2025, 235
  • [24] Green, Quantized Federated Learning Over Wireless Networks: An Energy-Efficient Design
    Kim, Minsu
    Saad, Walid
    Mozaffari, Mohammad
    Debbah, Merouane
    IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, 2024, 23 (02) : 1386 - 1402
  • [25] Energy-Efficient Dynamic Asynchronous Federated Learning in Mobile Edge Computing Networks
    Xu, Guozeng
    Li, Xiuhua
    Li, Hui
    Fan, Qilin
    Wang, Xiaofei
    Leung, Victor C. M.
    ICC 2023-IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS, 2023, : 160 - 165
  • [26] TOWARD ENERGY-EFFICIENT DISTRIBUTED FEDERATED LEARNING FOR 6G NETWORKS
    Khowaja, Sunder Ali
    Dev, Kapal
    Khowaja, Parus
    Bellavista, Paolo
    IEEE WIRELESS COMMUNICATIONS, 2021, 28 (06) : 34 - 40
  • [27] UAV-Enabled Energy-Efficient Aerial Computing: A Federated Deep Reinforcement Learning Approach
    Wu, Qianqian
    Liu, Qiang
    He, Ying
    Wu, Zefan
    IEEE TRANSACTIONS ON RELIABILITY, 2024,
  • [28] Joint Edge Association and Aggregation Frequency for Energy-Efficient Hierarchical Federated Learning by Deep Reinforcement Learning
    Ren, Yijing
    Wu, Changxiang
    So, Daniel K. C.
    ICC 2023-IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS, 2023, : 3639 - 3645
  • [29] Cooperative Energy-Efficient Management of Federated WiFi Networks
    Rossi, Claudio
    Casetti, Claudio
    Chiasserini, Carla-Fabiana
    Borgiattino, Carlo
    IEEE TRANSACTIONS ON MOBILE COMPUTING, 2015, 14 (11) : 2201 - 2215
  • [30] EFFICIENT AND RELIABLE OVERLAY NETWORKS FOR DECENTRALIZED FEDERATED LEARNING\ast
    Hua, Yifan
    Miller, Kevin
    Bertozzi, Andrea L.
    Qian, Chen
    Wang, Bao
    SIAM JOURNAL ON APPLIED MATHEMATICS, 2022, 82 (04) : 1558 - 1586