Occupancy-Gated Fuzzy Logic Control for Energy-Efficient Air-Conditioning in Context-Aware Smart Homes: A Simulation Study
Air-conditioning is the largest single electrical load in many homes, and the on/off thermostats that still govern most units run at full power, cycle the compressor often, and keep cooling empty rooms. This study designs and evaluates a context-aware fuzzy logic controller that reads two signals, room temperature and occupancy, and sets air-conditioner power. A single room was modelled as a first-order thermal plant and controlled by Mamdani fuzzy inference system. An initial design, V1, treated occupancy as a soft influence and overcooled empty rooms. A corrected design, V2, writes occupancy into the rule base as a gate with an explicit off state, so an empty room drives cooling to zero. The controllers were re-implemented in Python and compared against an on/off thermostat and an occupancy-aware thermostat under a fixed 180-minute test, a 24 h day, and 300 randomised scenarios, with switching counted for every controller by one threshold-crossing metric. With the room occupied throughout the fixed test, V2 stayed in the comfort band 99.2 % of the time on 1.05 kWh against 1.13 kWh for the thermostat, without switching. On realistic days V2 cut mean daily energy by 18.5 %, but occupied comfort fell to 88.7 % on average against 97.1 % for the thermostat, a trade-off of about eight points, and V2 still restarted the compressor about five times a day against 129 for the thermostat. Most of the saving came from not cooling an empty room. An occupancy-aware thermostat captured most of it without any fuzzy inference.
Authors
- Ikenna Augustine Onyegbadue (ORCID: https://orcid.org/0000-0002-6837-5509)
- Fred Abeibhode Izilein
- Chima Kesniel Igumbor
Institutions
- Igbinedion University (NG)
Publication Details
- Journal
- Lafia Journal of Scientific and industrial Research
- Published
- 2026-11-21
- DOI
- https://doi.org/10.62050/ljsir2027.v5n1.959
- Primary Topic
- Building Energy and Comfort Optimization
- Type
- article
- Field-Weighted Citation Impact
- 0.00