Application of Dempster-Shafer Theory in Sensor Data Fusion

Several machine learning algorithms and mathematical methods are being developed and implemented on various datasets available for indoor to detect occupancy for efficient utilization. In this chapter, environmental sensors of temperature, light, humidity, and CO2 datasets are used to analyze a room’s occupancy using the Dempster-Shafer theorem (DST) for better accuracy. DST is a mathematical theory to measure uncertainty in the hypothesis. It provides a combination rule to fuse evidence in better way to get the inference of an event. Visual Studio Code software tool is installed for 32-bit Windows 7 in laptops to program algorithms using pyds Python library. DST is applied to fuse different combinations of all four sensor data. With a cooperative data fusion of all four sensor data, an accuracy of 94.58% is obtained with the training dataset in the occupancy detection of a room.