Classification of Human Activities Indoors using Microclimate Sensors and Semiconductor Gas Sensors

Abstract

Nowadays, one of important problems faced by people in developed countries is poor indoor air quality (IAQ). Factors, which influence air inside buildings should be recognised for planning actions aimed at the improvement of indoor conditions. Our study was focused on human impact on IAQ. The aim of this work was the classification of the occurrence of occupants activities, which influence IAQ. The classification was based on measurements of indoor air using sensors. The presented analysis was focussed on the kind of sensors that are capable of providing the information which is most relevant for classification. Two groups of such devices were considered. The first included sensors which are typically used in microclimate measurements, i.e. temperature, relative humidity and CO2 concentration sensor. The second group included semiconductor gas sensors, which are considered as the sources of information about the chemical quality of indoor air. Classification tree was applied as the classifier. The obtained results showed that the measurement data provided by both groups of sensors can be applied for the classification of human activities, with the satisfactory performance. It may be understood that the impact of human activities on indoor air is very broad and may be examined using versatile sources of measurement data.