CoChat

Federated Learning for Internet of Things: A Comprehensive Survey

Open access
Nguyen, Dinh C., Ding, Ming, Pathirana, Pubudu N., Seneviratne, Aruna, Li, Jun, Vincent Poor, H. · 2021 · IEEE Communications Surveys & Tutorials

The DOI matches its Crossref record; 237 of the 237 references it cites were found in a registry or an index; 46 of the 198 sources its claims cite were checked for support.

Abstract

The Internet of Things (IoT) is penetrating many facets of our daily life with the proliferation of intelligent services and applications empowered by artificial intelligence (AI). Traditionally, AI techniques require centralized data collection and processing that may not be feasible in realistic application scenarios due to the high scalability of modern IoT networks and growing data privacy concerns. Federated Learning (FL) has emerged as a distributed collaborative AI approach that can enable many intelligent IoT applications, by allowing for AI training at distributed IoT devices without the need for data sharing. In this article, we provide a comprehensive survey of the emerging applications of FL in IoT networks, beginning from an introduction to the recent advances in FL and IoT to a discussion of their integration. Particularly, we explore and analyze the potential of FL for enabling a wide range of IoT services, including IoT data sharing, data offloading and caching, attack detection, localization, mobile crowdsensing, and IoT privacy and security. We then provide an extensive survey of the use of FL in various key IoT applications such as smart healthcare, smart transportation, Unmanned Aerial Vehicles (UAVs), smart cities, and smart industry. The important lessons learned from this review of the FL-IoT services and applications are also highlighted. We complete this survey by highlighting the current challenges and possible directions for future research in this booming area.

Overview

CoChat Summary

TL;DR

This survey argues that federated learning (FL) can make IoT systems more privacy-preserving, scalable, and communication-efficient by training models on distributed devices instead of centralizing raw data. It reviews FL in IoT services and applications, then summarizes lessons learned and open research challenges.

Key findings

  • FL is presented as an alternative to raw-data sharing in IoT: it enables “privacy-enhanced and scalable data sharing” and keeps raw data local while sharing learning results.
  • The survey identifies several IoT service areas where FL is useful: data sharing, data offloading and caching, attack detection, localization, mobile crowdsensing, and IoT privacy/security.
  • It also surveys FL in major application domains: smart healthcare, smart transportation, UAVs, smart city, and smart industry.
  • In the paper’s taxonomy tables, several reported results include:
  • IoT video data sharing with FL achieved 3.8% higher image-classification accuracy and 1.1% better object-detection performance over conventional FedAvg.
  • A UAV-swarm FL scheme reduced the convergence round by 35% compared with non-joint schemes.
  • A mobile crowdsensing FL scheme reported over 98% accuracy, with 23.9% runtime reduction and 33.3% communication-latency reduction for gradient aggregation.

Methods

  • This is a narrative survey, not an experimental study: it synthesizes prior FL and IoT literature and organizes it by service type, application domain, and FL architecture.
  • The paper’s structure includes an overview of FL and IoT, then sections on FL for IoT services and FL for IoT applications, followed by lessons learned and research challenges.
  • It also uses taxonomy tables to summarize prior work, including FL type, ML type, FL clients, FL aggregator, key contributions, and limitations.

Limitations

  • The paper itself states that existing work leaves several areas insufficiently covered, motivating this survey; it says prior surveys did not provide a comprehensive review of FL in IoT networks and applications.
  • It states that many FL-IoT problems remain open, including threats/security, communication and convergence issues, resource management, and the feasibility of deploying AI on constrained IoT sensors.
  • Evident from the text: this is a high-level survey with no new dataset, no controlled experiment, and no direct empirical validation of the reviewed methods by the authors themselves.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1109/comst.2021.3075439

Cites

237 of 237 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Organization of this article.
Fig. 1 Organization of this article.From the PDF
Image 2
From the PDF
Fig. 3 Types of FL models with data partitioning.
Fig. 3 Types of FL models with data partitioning.From the PDF
Fig. 4 Types of FL models with networking structure.
Fig. 4 Types of FL models with networking structure.From the PDF
Image 5
From the PDF
Fig. 6 The architecture of federated learning-based data sharing for IoV with blockchain [52].
Fig. 6 The architecture of federated learning-based data sharing for IoV with blockchain [52].From the PDF
Fig. 7 Federated attack detection and defense in FL-based IoT networks.
Fig. 7 Federated attack detection and defense in FL-based IoT networks.From the PDF
Fig. 8 Federated learning for blockchain-based crowdsensing in UAV networks.
Fig. 8 Federated learning for blockchain-based crowdsensing in UAV networks.From the PDF
Fig. 9 FL-IoT application domains.
Fig. 9 FL-IoT application domains.From the PDF
Fig. 10 Federated learning for smart healthcare: A case study for COVID-19 image classification with blockchain [128].
Fig. 10 Federated learning for smart healthcare: A case study for COVID-19 image classification with blockchain [128].From the PDF
Fig. 11 Federated learning for UAV-based ground air quality monitoring [147].
Fig. 11 Federated learning for UAV-based ground air quality monitoring [147].From the PDF
Fig. 12 Federated learning for cooperative industry [161].
Fig. 12 Federated learning for cooperative industry [161].From the PDF

Verification

Claims the paper makes

40 Valid · 7 Not valid · 135 Not checked · 63 Own results and interpretations

107 more claims · 22 with their cited sources checked

Create a free account to see all 245 claims and to verify the 59 references still open.

See all claims — free account
  • Probably not a substantial claim. 56
  • It may state two separate claims as one. 89
  • The quote only partly states this claim. 2

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Research works printed but not linked: 5

  1. 2020, Cisco Global Cloud Index Forecast and Methodology 2016–2021 White Paper
  2. xu, 2019, ELFISH Resource-aware federated learning on heterogeneous edge devices
  3. rodriguez-barroso, 2020, Dynamic federated learning model for identifying adversarial clients
  4. cai, 2020, TinyTL: Reduce memory, not parameters for efficient on-device learning, Proc Adv Neural Inf Process Syst, 33
  5. giust, 2017, Multi access edge computing overview of ETSI–IEEE future networks, IEEE 5G Tech Focus, 1, 4

Cited by

See all 1,507

Metadata

Authors
Nguyen, D. C., Ding, M., Pathirana, P. N., Seneviratne, A., Li, J., & Vincent Poor, H.
Affiliations
  • Commonwealth Scientific and Industrial Research Organisation (AU)
  • Data61 (AU)
  • Deakin University (AU)
  • UNSW Sydney (AU)
  • Nanjing University of Science and Technology (CN)
Venue
IEEE Communications Surveys & Tutorials
Published
2021
Volume
23
Issue
3
Pages
1622-1658
ISSN
  • 1553-877X (ISSN-L)
  • 2373-745X
OpenAlex
W3155912831
Cited by
1,503
References
171
Citations by year
2026: 325 · 2025: 458 · 2024: 359 · 2023: 243 · 2022: 106 · 2021: 16
FWCI
96.12
Type
journal article
Language
en
Open access
Locations
  • IEEE Communications Surveys & Tutorials (publishedVersion, cc-by)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University) (cc-by)
Funders
  • CSIRO Data61, Australia
  • U.S. National Science Foundation (CCF-1908308)
  • National Natural Science Foundation of China (61872184)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
  • Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
  • Physical Sciences › Computer Science › Computer Science Applications › Mobile Crowdsensing and Crowdsourcing
Keywords
federated learning, Internet of Things, distributed learning, data privacy, data offloading, caching, attack detection, localization

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.