A groundbreaking study from the University of the West of Scotland (UWS) has developed a new way to monitor people’s health without touching them — using ambient data from indoor room sensors.
Researchers, PhD student Cezar Anicai and Professor Muhammad Zeeshan Shakir, believe this non-invasive system could offer a privacy-respecting alternative to wearable devices and medical sensors, especially in care settings.
600 minutes of data under varied indoor conditions from the study’s 14 participants was compiled, with variables such as temperature, humidity, light, sound, pressure, air quality, and physiological responses measured to assess the wellbeing of participants.
Cezar Anicai, a PhD student at UWS said:
“We know that the quality of our indoor environments – temperature, air, light, noise – has a direct effect on health outcomes.
“However, traditional wearable or medical-grade sensors, while accurate, are intrusive and unsuitable for long-term, continuous use. Moreover, they do not offer any information about what is influencing the monitored signals.
“This research demonstrates how smart, indoor spaces equipped with ambient sensors can monitor critical physiological signals, and it enables us to measure the connection between ambient conditions and health indicators with scientific precision.
“We believe this technology would be especially useful in settings such as care homes for the elderly, where residents are largely indoors and often must be closely monitored. This technology would allow staff to monitor the health of their residents less intrusively and distressing for residents.”
The implications of this research could be far-reaching, offering potential applications in areas such as occupational health, lifestyle modification, and predictive health monitoring.
Professor Muhammad Zeeshan Shakir said:
“The machine learning models Cezar developed show robust performance in estimating cardiac and electrodermal activity from ambient signals, offering a non-invasive alternative to wearable health sensors, in a way that respects users’ privacy. This represents a paradigm shift in ambient health sensing, enabling real-time, continuous monitoring without physical contact.”
Following on from the initial study, the team want to further expand the dataset by including more participants, a wider variety of indoor settings, additional environmental and physiological signals. In parallel, they hope to develop more advanced machine learning models that can adapt quickly to new users to allow for fast and seamless onboarding.
The full dataset and accompanying study are available in the May 22, 2025, issue of Scientific Data. Click here for more information or to access the dataset.
This research aligns with the United Nations Sustainable Development Goals (UNSDGs), specifically Goal 3: Good Health and Well-Being.