Conference Paper

Monitoring daily human activity has been of considerable relevance in diagnosing an individual's physical and cognitive health conditions. This type of monitoring has gained even more importance with the ongoing digital transformation in the healthcare sector, particularly in ambient assisted living (AAL) environments, where continuous and non-intrusive monitoring is crucial for elderly care and early health diagnosis. One promising approach is activity recognition using air quality data, considering that many routine tasks affect indoor air parameters. This paper presents the designed and implemented portable Internet of Things (IoT) air quality monitoring and explores its capability of recognizing indoor activities using threshold-based analysis and machine learning algorithms. The study focuses on detecting common and indispensable activities, such as meal preparation, personal hygiene, and the occupancy of a living room. Results show that the proposed system and its machine learning model achieve high accuracy, exceeding 97%, in recognizing these activities based solely on indoor air quality measurements.