Our attention was focused on the identification of activities affecting air quality, which occur in quick-service restaurants (QSR). The work was based on a measurement study of selected kebab stores in the Polish city of Wrocław. It demonstrated that activities taking place in kebabs altered air quality. The associated changes in air parameters such as temperature, relative humidity, CO2 concentration, and the content of volatile organic compounds could be detected by utilizing a simple, multi-sensor device. In the measurement data, there were identified multidimensional patterns, which proved to be specific for the following categories of activities: Night Hours, Outlet Preparation, Food Preparation, Operation Hours, and Cleaning. Their occurrence was recognized by pattern recognition methods with a true positive rate greater than 99%. We demonstrated that the recognition may be based on measurements performed in various locations within the kebab store. Although patterns of the individual categories of activities largely varied between kebab stores, a similar performance of recognition was achieved for all restaurants. The obtained results entitled us to conclude that it is possible to detect activities of QSR, which influenced air quality, with the application of sensor technique and pattern recognition. The proposed approach may be applied to this type of object in general.
Keywords: indoor air quality; pattern recognition; sensing.