Propagation of the Measurement Uncertainty of Wearable Sensors for Thermal Comfort Assessment

The development of personal comfort models (PCMs) is pivotal for the optimization of both the occupants' comfort and the building energy consumption; wearable sensors assessing physiological parameters can be exploited for this aim. This paper considers commercial and laboratory prototypes of wearable sensors and aims at evaluating the propagation of the input measurement uncertainties on the features computed for the development of PCMs. Different types of wearables, suitable for physiological monitoring in a living environment, are considered: smartwatch, smartband, smart garment, smart ring, patch, and headband. The Monte Carlo simulation method is exploited for the uncertainty analysis. The results show that some features are more robust than others and this is relevant when selecting a feature subset for a model creation. The findings are useful for the design of PCMs aiming at improving the quality of life in indoor living environments, developed with a human-centric view.