Vision-based thermal comfort quantification for HVAC control

摘要: Abstract This study presents a vision-based approach that employs RGB video images as the sole source for inferring thermoregulation states in human body response to thermal condition variations indoor environments. The primary objective is contribute our envisioned thermoregulation-based HVAC control leverages actual demands from end-users’ increased energy efficiency. Given system calls measurement techniques under four constraints of non-intrusiveness, applicability, sensitivity, and ubiquity (i.e., feasibility scalability), this investigated potentials ubiquitously obtainable (through webcams or smartphones). Using photoplethysmography (PPG), well-known optical technique measuring blood volume changes microvascular bed skin, we have leveraged mechanism flow skin surface (blood vessels' dilation constriction) heat dissipation regulations, reflected PPG signal's amplitude. subtle signals their susceptibility noise, proposed framework uses combination independent component analysis adaptive filtering reduce unwanted in-band artifacts while preserving amplitude information indicates thermophysiological states. was experimentally evaluated using transient conditions account applicability sensitivity attributes. Therefore, without considering an acclimation time stabilized states, subjects were exposed varying temperatures (∼20–30 °C) reporting sensations. In total, 10 out 15, positive correlation between indicators, temperature, sensations observed demonstrating promising potential occupants with sufficient sensitivity.