Office Occupancy Detection based on Power Meters and BLE Beaconing

Azkario Rizky Pratama


468 Downloads (Pure)


Energy consumption for both residential and non-residential buildings is significant and has been increasing regularly. For non-residential buildings, asking the user to be directly involved in energy saving can be challenging as occupants (e.g., employees) are less aware of and affected by high energy bills compared to their domestic situation. Employees are less careful when leaving empty office spaces heated and illuminated, resulting in unnecessary energy consumption. This thesis focuses on finding solutions for solving energy waste in non-residential buildings by automatically detecting the presence, thus enabling energy-saving automation.

To reduce energy consumption due to unnecessary use, precise and detailed user contexts play an important role. User contexts (e.g., occupancy and activity of users) provide grounds to buildings’ control and energy management systems for efficient lighting and HVAC actuation. We explore sensing systems that indicate occupancy. Namely, we extract occupancy from power consumption (i.e., power metering or sub-metering systems) and proximity location (i.e., mobile phones with beaconing systems). We investigate several strategies and machine learning algorithms to infer occupancy from these sources. We also study fusions at the decision-level and feature-level. The former allows sub-systems to infer local decisions and finally combines the outputs to form a final decision. The latter yields only decision after sensor readings have been combined. The approaches are tested in actual office environments populated by researchers and software developers. We finally discuss potential energy saving, user privacy, and portability to provide insight into how the proposed occupancy detection systems may impact building use and control.
Originele taal-2English
KwalificatieDoctor of Philosophy
Toekennende instantie
  • Rijksuniversiteit Groningen
  • Lazovik, Alexander, Supervisor
  • Aiello, Marco, Supervisor
Datum van toekenning18-dec.-2020
Plaats van publicatie[Groningen]
StatusPublished - 2020


Duik in de onderzoeksthema's van 'Office Occupancy Detection based on Power Meters and BLE Beaconing'. Samen vormen ze een unieke vingerafdruk.

Citeer dit