Internet of Things. User-Centric IoT. First International Summit, IoT360 2014, Rome, Italy, October 27-28, 2014, Revised Selected Papers, Part I

Research Article

A Learning Approach for Energy Efficiency Optimization by Occupancy Detection

Download
354 downloads
  • @INPROCEEDINGS{10.1007/978-3-319-19656-5_2,
        author={Vitor Mansur and Paulo Carreira and Artur Arsenio},
        title={A Learning Approach for Energy Efficiency Optimization by Occupancy Detection},
        proceedings={Internet of Things. User-Centric IoT. First International Summit, IoT360 2014, Rome, Italy, October 27-28, 2014, Revised Selected Papers, Part I},
        proceedings_a={IOT360},
        year={2015},
        month={7},
        keywords={Energy savings Users preferences Occupancy detection HVAC system Thermal comfort},
        doi={10.1007/978-3-319-19656-5_2}
    }
    
  • Vitor Mansur
    Paulo Carreira
    Artur Arsenio
    Year: 2015
    A Learning Approach for Energy Efficiency Optimization by Occupancy Detection
    IOT360
    Springer
    DOI: 10.1007/978-3-319-19656-5_2
Vitor Mansur1,*, Paulo Carreira,*, Artur Arsenio,*
  • 1: Universidade de Lisboa
*Contact email: vitor.mansur@ist.utl.pt, paulo.carreira@ist.utl.pt, arsenio@alum.mit.edu

Abstract

Building Automation Systems control HVAC systems aiming at optimizing energy efficiency and comfort. However, these systems use pre-set configurations, which usually do not correspond to occupants’ preferences. Although existing systems take into account the number of occupants and the energy consumption, individual occupant preferences are disregarded. Indeed, there is no way for occupants to specify their preferences to HVAC system. This paper proposes an innovation in the management of HVAC systems: a system that tracks the occupants preferences, and manages automatically the ventilation and heating levels accordingly to their preferences, allowing the system to pool its resources to saving energy while maintaining user comfort levels. A prototype solution implementation is described and evaluated by simulation using occupants’ votes. Our findings indicate that one of the algorithms is able to successfully maintain the appropriate comfort levels while also reducing energy consumption by comparing with a standard scenario.