IoT-Based Sensor Data Fusion for Occupancy Sensing Using Dempster-Shafer Evidence Theory for Smart Buildings

With the advent of Internet of Things technologies and smart buildings, the need for building automation systems that automatically performs computations without the intervention of humans have also increased. This paper deals with detection of occupancy in a room from various ambient sources like temperature, humidity, light, and CO2. With the help of this system, remote monitoring of the building as well as leveraging control on the indoor parameters through HVAC control systems is possible at real-time. This paper adopts Dempster-Shafer evidence theory for fusing sensory information collected from heterogeneous sensors, assigns probability mass assignments (PMAs) to the raw sensor readings, and finally performs mass combination to derive a conclusion about the occupancy status in a room. A PMA function has been proposed for this purpose. The results reveal a substantially high percentage of accuracy (up to 99.09%) which was observed to increase with the increase in number of fusion parameters.