Federated Learning for DL-CSI Prediction in FDD Massive MIMO Systems

dc.authoridElbir, Ahmet M./0000-0003-4060-3781
dc.authoridOhtsuki, Tomoaki/0000-0003-3961-1426
dc.authoridHou, Weihao/0000-0002-9944-6402
dc.authorwosidElbir, Ahmet M./X-3731-2019
dc.authorwosidGui, Guan/AAG-3593-2019
dc.contributor.authorHou, Weihao
dc.contributor.authorSun, Jinlong
dc.contributor.authorGui, Guan
dc.contributor.authorOhtsuki, Tomoaki
dc.contributor.authorElbir, Ahmet M.
dc.contributor.authorGacanin, Haris
dc.contributor.authorSari, Hikmet
dc.date.accessioned2021-12-01T18:50:11Z
dc.date.available2021-12-01T18:50:11Z
dc.date.issued2021
dc.department[Belirlenecek]en_US
dc.description.abstractIn frequency-division duplexing (FDD) massive multiple-input multiple-output (MIMO) systems, deep learning for predicting the downlink channel state information (DL-CSI) has been extensively studied. However, in some small cellular base stations (SBSs), a small amount of training data is insufficient to produce an excellent model for CSI prediction. Traditional centralized learning (CL) based method brings all the data together for training, which can lead to overwhelming communication overheads. In this work, we introduce a federated learning (FL) based framework for DL-CSI prediction, where the global model is trained at the macro base station (MBS) by collecting the local models from the edge SBSs. We propose a novel model aggregation algorithm, which updates the global model twice by considering the local model weights and the local gradients, respectively. Numerical simulations show that the proposed aggregation algorithm can make the global model converge faster and perform better than the compared schemes. The performance of the FL architecture is close to that of the CL-based method, and the transmission overheads are much fewer.en_US
dc.description.sponsorshipJSPS KAKENHIMinistry of Education, Culture, Sports, Science and Technology, Japan (MEXT)Japan Society for the Promotion of ScienceGrants-in-Aid for Scientific Research (KAKENHI) [JP19H02142]; Ministry of Industry and Information Technology of China [TC190A3WZ-2]; National Natural Science Foundation of ChinaNational Natural Science Foundation of China (NSFC) [61901228]; Summit of the Six Top Talents Program of Jiangsu [XYDXX010]; Program for High-Level Entrepreneurial and Innovative Team [CZ002SC19001]; Key Laboratory of Universal Wireless Communications (BUPT) of Ministry of Education of ChinaMinistry of Education, China [KFKT-2020106]en_US
dc.description.sponsorshipThis work was supported in part by the JSPS KAKENHI under Grant JP19H02142; in part by the Major Project of the Ministry of Industry and Information Technology of China under Grant TC190A3WZ-2; in part by the National Natural Science Foundation of China under Grant 61901228; in part by the Summit of the Six Top Talents Program of Jiangsu under Grant XYDXX010; in part by the Program for High-Level Entrepreneurial and Innovative Team under Grant CZ002SC19001; and in part by the Project of the Key Laboratory of Universal Wireless Communications (BUPT) of Ministry of Education of China under Grant KFKT-2020106. The associate editor coordinating the review of this article and approving it for publication was F. Tariq.en_US
dc.identifier.doi10.1109/LWC.2021.3081695
dc.identifier.endpage1814en_US
dc.identifier.issn2162-2337
dc.identifier.issn2162-2345
dc.identifier.issue8en_US
dc.identifier.scopus2-s2.0-85107214434en_US
dc.identifier.scopusqualityQ1en_US
dc.identifier.startpage1810en_US
dc.identifier.urihttps://doi.org/10.1109/LWC.2021.3081695
dc.identifier.urihttps://hdl.handle.net/20.500.12684/10840
dc.identifier.volume10en_US
dc.identifier.wosWOS:000682125800045en_US
dc.identifier.wosqualityQ1en_US
dc.indekslendigikaynakWeb of Scienceen_US
dc.indekslendigikaynakScopusen_US
dc.language.isoenen_US
dc.publisherIeee-Inst Electrical Electronics Engineers Incen_US
dc.relation.ispartofIeee Wireless Communications Lettersen_US
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanıen_US
dc.rightsinfo:eu-repo/semantics/closedAccessen_US
dc.subjectChannel state informationen_US
dc.subjectcentralized learningen_US
dc.subjectfederated learningen_US
dc.subjectsmall cellular base stationen_US
dc.subjectmacro base stationen_US
dc.titleFederated Learning for DL-CSI Prediction in FDD Massive MIMO Systemsen_US
dc.typeArticleen_US

Dosyalar

Orijinal paket
Listeleniyor 1 - 1 / 1
Küçük Resim Yok
İsim:
10840.pdf
Boyut:
666.99 KB
Biçim:
Adobe Portable Document Format
Açıklama:
Tam Metin / Full Text