Enhancing personal comfort - A machine learning approach using physiological and environmental signals

Aug 5, 2026

The assessment of occupants’ thermal sensation (TS) supports wellbeing and helps optimize building energy consumption. Machine learning can predict TS using physiological and environmental parameters, yet selecting the best feature subset is essential. This study analyzes the correlation between physiological parameters and TS and identifies optimal features for ML-based prediction. An experimental campaign collected signals via wearable sensors, while TS was gathered through a dedicated questionnaire. Results highlight strong physiological feature influence, with classifiers reaching about 90% accuracy, enabling personalized comfort systems that improve comfort and energy efficiency via a human-centric approach.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Introduction

Thermal sensation and wellbeing in buildings

Personal Comfort Models and limitations of conventional methods

Survey-based vs sensor-based assessments

Predictive Mean Vote (PMV) and group-based comfort models

Measurement 217 (2023) 113047 Contents lists available at ScienceDirect Measurement journal [homepage:](homepage: www.elsevier.com/locate/measurement)[ www.elsevier.com/locate/measurement](homepage: www.elsevier.com/locate/measurement) | Enhancing personal comfort: A machine learning approach using physiological and environmental signals measurements | | | | | --- | --- | --- | --- | | Gloria Cosolia, *, Silvia Angela Mansib, Ilaria Pigliautilec, Anna Laura Pisello c, d, Gian Marco Revela, Marco Arnesanob
a Department of Industrial Engineering and Mathematical Sciences, Universit`a Politecnica delle Marche, Ancona, Italy b Universit`a Telematica eCampus, Novedrate, Italy
c Department of Engineering, Department of Excellence 2023-2027, via G. Duranti 93, Universit`a degli Studi di Perugia, Italy d Centro interuniversitario sull’inquinamento e l’ambiente M. Felli, via G. Duranti 67, Universita` degli Studi di Perugia, Italy | | | | | A R T I C L E I N F O | | A B S T R A C T | | | Keywords:
Thermal sensation Personal comfort model Indoor comfort Wearable sensors Physiological signals Machine learning Surveys | | The assessment of the occupants’ thermal sensation (TS) in a living environment is fundamental to enhance wellbeing and optimize building energy consumption. Machine Learning (ML)-based approaches can be adopted for TS prediction exploiting physiological and environmental parameters, but identifying an optimal features subset is fundamental. This work aims at assessing the correlation between physiological parameters and TS, hence selecting the optimal feature subset for ML-based TS prediction. A dedicated experimental campaign was designed to gather signals through wearable sensors; the actual TS was collected via a specific questionnaire. The results prove the weight of physiological features on the TS determination; ML classifiers achieved an accuracy of up to ≈90% by using physiological and environmental parameters. The strategic potential of personalized comfort systems enables the optimization of both comfort and energy efficiency of a building according to a human-centric approach. | | 1. Introduction The measurement of thermal sensation (TS) can provide very relevant information on buildings occupants’ well-being; in fact, the subjective perception of thermal conditions leads to changes in vital parameters, hence modifying the human’s status and reflecting into her/ his comfort and, more in general, well-being. Thermal comfort is a crucial aspect of the built environment, due to its effect on the occupants’ health [1], well-being, and work efficiency [2], as well as on buildings energy consumption [3]. This can be observed from a twofold perspective: on the one side, a subject’s physiological status is affected (among others) by indoor thermal conditions, hence changes in thermal perception of the surrounding environment and the related TS are correlated with changes in physiological responses [4]; on the other side, the individual response to certain ambient conditions can be interpreted and exploited for control purposes, in particular in the context of Personal Comfort Models (PCMs) [5,6]. At present, PCMs are being studied to actually take account of the real needs of a building occupant, and they generally outperform conventional models [7,8]. For example, Williamson et al. [9] investigated the relationship between thermal comfort and well-being of elderly occupants in order to provide some recommendations for policymakers as well as guidelines for living environments design, given that the built environment affects also the health outcomes [10]. Moreover, it should be considered that such models can be beneficial to optimize the buildings energy demand (with energy savings from 4% to 60%[11]), which has a huge impact on the environment (the building and construction sector accounts for the 36% of the total energy consumption and for the 37% of energy-related CO2 emissions [3]

FAQs

Why is thermal sensation (TS) important in living environments?

TS is tied to occupants’ wellbeing and can influence health and work efficiency. It also relates to building energy consumption through how people perceive indoor thermal conditions.

How does this study predict thermal sensation using machine learning?

It correlates physiological parameters with TS to select an optimal feature subset for ML-based TS prediction. Wearable sensors collect physiological and contextual signals, while TS is measured through a questionnaire.

What do the results show about the role of physiological features?

Physiological features significantly affect TS determination. ML classifiers achieve up to about 90% accuracy when using both physiological and environmental parameters.

Aug 5, 2026

Upload by Lucas Martin