A semantic event approach to enrich device-free indoor localization data

Device-free indoor localization technologies can achieve fine-grained localization and have the advantage of being non-intrusive. However, interpreting the low-level data produced by these technologies can be challenging. In this paper, we investigate the application of stream reasoning technologies to enrich and exploit such data. In particular, we are interested in the semantic fusion of events coming from various ambient sensors (door sensors, smart plugs, motion sensors ... ), including a ceiling-mounted Pyroelectric InfraRed (PIR) sensor array that provides accurate multi-user positioning on a 2D plane. We show how existing stream reasoning technologies can be used to detect complex events, such as the movements and actions of multiple users in a real environment. These events represent useful information which can later be used for human-centered use cases, such as detecting activities or abnormal user behavior.