Machine Learning Based Occupancy Detection in Smart Environments: A Residential Flat Case

With the increasing spread of smart cities and smart building technologies, the concept of the Internet of Things (IoT) has emerged, and their control has also gained importance. Automatic control of devices in apartments in a smart building is important in terms of energy saving and comfort. In this context, this study focuses on detecting the occupancy in order to automatically control the devices in a smart home. To achieve this, measurements of inorganic gas, temperature, pressure, humidity and light levels in the residential flat are used. An original dataset of approximately 5 days duration is collected with the developed sensor device, obtained data is processed with various machine learning methods and occupancy detection is carried out.