A Physiological-Signal-Based Thermal Sensation Model for Indoor Environment Thermal Comfort Evaluation

Area of Science:

• Building Science
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Traditional heating, ventilation, and air conditioning (HVAC) systems use static models like Predicted Mean Vote (PMV) for thermal comfort.
• These models do not fully capture individual human thermal comfort due to limitations in considering dynamic physiological responses.
• Advancements in sensor technology enable the integration of physiological data into comfort models.

Purpose of the Study:

• To develop and validate a novel thermal sensation (TS) model incorporating human physiological signals alongside environmental parameters.
• To compare the performance of the proposed physiological-signal-based TS model against the traditional PMV model.
• To identify the most influential physiological signals for accurate thermal sensation prediction.

Main Methods:

• Climate chamber experiments were conducted with young subjects under controlled environmental conditions (temperature, humidity, fan speed).
• Physiological data including electrocardiogram (ECG), electroencephalogram (EEG), electromyogram (EMG), galvanic skin response (GSR), and body temperatures were collected.
• A new TS model was developed using these physiological signals and environmental data, and its accuracy was compared to the PMV model using RMSE and R² metrics.

Main Results:

• The physiological-signal-based TS model demonstrated superior performance compared to the PMV model, with lower Root Mean Square Error (RMSE) (0.75 vs. 1.07) and higher R-squared (R²) (0.77 vs. 0.43).
• Electromyogram (EMG), body temperature, ECG, and EEG were identified as the most significant physiological signals, in descending order of importance.
• The findings indicate higher accuracy and better explainability for the proposed physiological-signal-based model.

Conclusions:

• Physiological signals significantly enhance the accuracy and explainability of thermal sensation prediction models.
• The developed TS model offers a more personalized and accurate approach to indoor environmental control compared to traditional methods.
• Further research is recommended to explore broader applications, including diverse age groups, health conditions, and activity levels (static, motion, sports).