Creating and sharing knowledge for telecommunications

Unobtrusive Cardio-Respiratory Assessment for Different Indoor Environmental Conditions

Rodrigues, M ; Postolache, O. ; Cercas, Cercas, F.

IEEE Sensors Journal Vol. 22, Nº 23, pp. 23243 - 23257, December, 2022.

ISSN (print): 1530-437X
ISSN (online):

Scimago Journal Ranking: 0,99 (in 2022)

Digital Object Identifier: 10.1109/JSEN.2022.3207522

Download Full text PDF ( 3 MBs)

Downloaded 4 times

Abstract
The achievement of a healthy indoor environ- ment—an indispensable requirement for human well-being—relies on the establishment of good air quality conditions and thermal comfort levels. This study addresses the development of a healthcare-focused Internet-of-Things (IoT) system for the cardiac assessment using photoplethysmography (PPG) and ballistocardiography (BCG) that can be applied in any indoor environment. Results from its experimental application include the analysis of heart rate variability (HRV) and respiratory status from a group of young adults under different thermal conditions and relative humidity (RH) levels. The final goal was to investigate the effects of short-term exposure to different air temperature and humidity conditions on human thermal comfort based on the analysis of the autonomic nervous system (ANS) activity. The experiment took place in a regular and nonisothermal office environment. Simulations based on computational fluid dynamics (CFD) were conducted to estimate heat distribution and select the optimal locations for the environmental sensor nodes’ placement. Results indicated that activation of the parasympathetic system was more notable in short-term exposures to environments with different air temperatures (24 °C–30 °C) than with different levels of RH (50%–70%). Changes in ambient air temperature led to the activation of thermal regulatory reflexes and thermal discomfort, perceived by an increase in the ratio of low frequency to high frequency (LF/HF) at higher air temperature exposures. In addition, HRV parameters were used to train machine learning models on predicting if the user is in a thermally comfortable environment or a discomfortable environment (hot or humid), having achieved an 87% average accuracy.