Energy-efficient federated learning methods for IoT environment
Amutha Prabakar Muniyandi, Feslin Anish Mon, L. Godlin Atlas, Mahmoud Ahmad Al‐Khasawneh · 2024
The rapid expansion of the Internet of Things (IoT) has led to an unprecedented surge in data generation and processing demands. Federated Learning (FL) has emerged as a promising framework for enabling collaborative model training across distributed IoT devices while preserving data privacy. Given the resource-constrained nature of IoT devices, developing energy-efficient FL techniques is crucial to ensure sustainability and prolonged device operation. This chapter provides a comprehensive exploration of cutting-edge energy-efficient methods tailored for FL in IoT environments. Key contributions include an in-depth analysis of energy consumption challenges specific to IoT-driven FL and the introduction of innovative strategies to optimize communication and computation overheads. The chapter systematically reviews and evaluates techniques such as model compression, adaptive aggregation, and participant selection. Furthermore, it examines the role of energy-aware resource allocation and edge computing in enhancing the efficiency of Federated Learning. Empirical studies and simulations are presented to demonstrate the feasibility and performance of these methods, offering valuable insights into their practical deployment. This chapter serves as a comprehensive guide for researchers and practitioners aiming to leverage Federated Learning in IoT ecosystems, fostering innovation in sustainable and intelligent IoT solutions by addressing the critical intersection of FL and energy efficiency.