Tools
Tools
Kochar, Akshay (2025) Federated Learning and Edge Computing for Latency Reduction in Smart Farming IoT Sensor Data Filtering. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration Manual)
Download (1MB) | Preview |
Abstract
The integration of Federated Learning and Edge Computing has emerged as a promising approach to handle latency, security, and bandwidth limitations in IoT sensor networks for smart farming. Traditional cloud-based models are usually afflicted with high communication costs and privacy concerns, hence proven to be inefficient for real-time agricultural applications. FL enables decentralized machine learning, thus allowing models to be trained right on IoT devices without requiring any raw data centralization, hence preserving data privacy and reducing transmission overhead. On the other hand, Edge Computing enhances local processing of data for reduced latency, reduced dependency on cloud infrastructure for quicker decision-making. However, despite advantages in both, certain key challenges like heterogeneous data processing, communication overhead, resource constraints, and security vulnerabilities remain concerning. This paper reviews the related literature with regard to the existing FL, Edge Computing, and IoT data filtering techniques. It outlines the critical research gap on the scalability in large-scale farms, energy-efficient learning models, and secure FL. The study explores enhancing the privacy, efficiency, and real-world deployment issues of FL techniques within an agricultural IoT system. This research hence tries to bridge the gaps with respect to the optimization of FL in smart farming for reducing latency by filtering data in real time for decision-making with increased security concerning IoT-driven agricultural systems.
Actions (login required)
 |
View Item |