Personalised Environmental Regulation for IoT Based on Perceived User Comfort

Recent works on home and office automation are either limited to specific tasks, or require significant user intervention. Furthermore, most solutions are based on the behavior of a typical user, and not each individual. This work proposes an office automation system that learns the preferences of e...

Full description

Saved in:
Bibliographic Details
Published in2024 Global Information Infrastructure and Networking Symposium (GIIS) pp. 1 - 6
Main Authors Contreras, Guillermo Ponce, de Azevedo Barbosa, Gustavo Guedes, Caitite, Vitor Gabriel Reis, Macedo, Daniel F., Veloso, Adriano Alonso
Format Conference Proceeding
LanguageEnglish
Published IEEE 19.02.2024
Subjects
Actuators
Behavioral sciences
Control
Internet of Things
Machine learning
Office automation
Personalisation
Regulation
Task analysis
Online AccessGet full text

Cover

More Information
Summary:Recent works on home and office automation are either limited to specific tasks, or require significant user intervention. Furthermore, most solutions are based on the behavior of a typical user, and not each individual. This work proposes an office automation system that learns the preferences of each person using machine learning. The model is individualized, and can regulate different actuators since its outputs are independent of the type of sensor inputs. An experimental analysis over a demonstrator indicates that each user has different notions of comfort. Further, the user comfort increased by up to 87% when compared to an uncontrolled room, and the user intervened manually in the environment un to 47% less times.
ISSN:2150-329X
DOI:10.1109/GIIS59465.2024.10449726