Announcement of data collection in Väre building: Indoor occupancy detection using machine learning methods and CO2 sensors

In some of Otaniementie 14 rooms, data on the number of occupants will be collected for research purposes from 15.4.2024 to 27.9.2024. This data is gathered with machine vision sensors. The sensors do not record images and do not identify individual persons. The sensors only record the quantity of people entering and exiting the room. A notification of the data collection will also be placed near the entrances of the rooms included in the study. The rooms included in the study are all meeting rooms.

The collected data will be used in a study where the occupancy of the room is modeled using a carbon dioxide sensor and machine learning. The carbon dioxide sensor can estimate the number of people in the room. However, to train a machine learning model for this purpose and to check the model's accuracy, an accurate count of occupants in the space is needed. In this case, it is obtained through an occupant counting sensor. If the number of occupants can be estimated anonymously using the carbon dioxide sensor data, it can be used, among other things, to save energy in air conditioning and heating.