Autor(es):
Ventura, Pedro ; Khodamoradi, Mohammad ; Costa, Ruben ; Figueiras, Paulo ; Jardim-Gonçalves, Ricardo
Data: 2025
Identificador Persistente: http://hdl.handle.net/10362/199715
Origem: Repositório Institucional da UNL
Assunto(s): Ambient Quality; Federated Learning; IoT; Machine Learning; General Computer Science; Ambient Quality; Ambient Quality; Federated Learning; Federated Learning; IoT; IoT; Machine Learning; Machine Learning; General Computer Science; General Computer Science
Descrição
Nowadays, modern office environments require the maintenance of optimal working conditions to enhance employee productivity and well-being. Traditional centralized data processing systems present challenges such as data privacy concerns, high communication latency, and inefficient use of network resources. Addressing these challenges is essential for improving office space efficiency and comfort. FL offers an innovative solution, by decentralizing data processing, allowing it to occur locally on edge devices. This approach provides several advantages over traditional centralized methods, including enhanced data privacy, reduced communication latency, and more efficient use of network resources. The proposed solution involves implementing FL models in Internet of Things (IoT) device systems within office buildings. Key environmental parameters measured by sensors include illuminance, air quality, temperature, and noise levels. These measurements are crucial for optimizing the office environment, thus improving operational efficiency and employee satisfaction.