This paper is available on arxiv under CC 4.0 license.
Authors:
(1) Berrenur Saylam;
(2) Ozlem Durmaz ¨ ˙Incel.
Table of Links
Federated Learning: Collaborative and Privacy-Preserving Model Training
State of the Art Federated Learning Software Platforms and Testbeds
Sensors and Edge-Sensing Devices
Federated Learning Applications on Sensing Devices
Conclusions, Acknowledgment, and References
Highlights
• Cutting-Edge Perspective: Our manuscript, titled “Federated Learning on Edge Sensing Devices: A Review,” provides an up-to-date and comprehensive exploration of federated learning (FL) within the context of edge sensing devices. It offers a fresh perspective on this rapidly evolving field, showcasing its relevance to the contemporary landscape of data analysis and privacy preservation.
• Addressing Critical Challenges: We address the pressing challenges associated with traditional centralized data analysis methods, such as privacy concerns, hardware constraints, and connectivity limitations. By highlighting the limitations of conventional approaches, our research underscores the urgency of adopting FL techniques to unlock the full potential of edge sensing devices.
• Wide-Ranging Application Domains: The manuscript sheds light on the broad spectrum of application areas where edge sensing devices play a pivotal role, including healthcare, environmental monitoring, automotive technology, industrial processes, aerospace, and agriculture. Our work emphasizes how FL can be a game-changer in extracting meaningful insights from sensor data in these critical domains.
• Comprehensive Review: We offer a comprehensive review of FL strategies tailored specifically for edge-sensing devices. This review encompasses the fundamental principles of FL, software frameworks for implementation, and practical testbeds. It serves as a valuable resource for researchers and practitioners seeking to navigate the complex landscape of FL in edge computing.
• Current Sensor Technologies: Our manuscript dives into the current state of sensor technologies, providing an in-depth exploration of the properties and capabilities of these devices. This insight is crucial for understanding the foundations upon which FL operates and how it can be effectively integrated into sensor-driven applications.
• Future Research Directions: By identifying open issues and charting future research directions, our work contributes to the ongoing discourse in the field. We invite further investigation into the potential of FL in edge sensing devices, offering a roadmap for researchers to explore uncharted territories and address emerging challenges.
• Timely and Substantial Contribution: Given the increasing importance of FL in the era of edge computing, our manuscript offers a timely and substantial contribution to the journal’s mission of advancing knowledge in the realms of diagnostics, human health, well-being, and activity recognition with wearable sensors.
• Unpublished and Unbiased: We assure the journal that neither this manuscript nor any of its contents are under consideration or published in another journal, demonstrating our commitment to presenting original and unbiased research.
This paper is