Federated learning for energy management applications

Remote monitoring systems play a crucial role in analyzing environmental dynamics across various smart industrial applications, including occupational health and safety, as well as environmental monitoring. In the context of industrial Internet of Things (IoT) systems, the vast proliferation of devices combined with the stringent performance requirements imposes significant pressure on resources such as computational power, network bandwidth, and device energy. The challenge is particularly pronounced when it comes to the distributed training of machine learning (ML) and deep learning (DL) models, which are essential for enabling intelligent functionalities in industrial IoT applications. These tasks are especially difficult to execute on resource-constrained devices operating within heterogeneous wireless networks. Managing these limitations while ensuring the effective deployment of ML/DL models requires innovative approaches to optimize resource utilization, energy management, and maintain system performance. In this chapter, we develop a framework that enhances energy efficiency in Federated Learning (FL) over wireless networks considering various network architectures and assumptions, and integrating energy-harvesting technologies.