A Performance Study of Cutting-Edge Technologies for Energy Consumption Prediction in Smart Buildings

Energy efficiency is crucial in modern smart building management. Effective energy management not only reduces operational costs but also promotes environmental sustainability by reducing carbon emissions. Additionally, optimizing energy usage improves occupant comfort and productivity, contributing to a healthier and more sustainable built environment. Recently researchers have focused on cutting-edge technologies to develop efficient models that predict energy consumption and ensure a tradeoff between occupant, provider and environment needs. Among such technologies, Internet of Things (IoT), edge/fog computing, and federated Learning (FL) have significantly have proven their efficiency in this domains. In this paper, we provide a performance analysis of such technologies for energy consumption prediction in smart buildings. According to a set of defined criteria, we select some recent proposed techniques and we study their performance through various assessment metrics. Our idea behind such comparison is to identify promising techniques while using Friedman test and furthermore highlight further open research problems in the domain. Real-world data has been used to measure and evaluate each approach providing valuable insights for practical implementation and deployment in smart building environments.