A Soft Real-Time Stream Reasoning Service for the Internet of Things

摘要

Most Internet of Things applications have their algorithms driven by elementary events or basic sensor data. While this may be sufficient for some applications, it precludes the possibility of describing and detecting semantic relationships among these events/data and to infer new application-relevant facts out of these relations, which could give an IoT application means of reasoning about the ambient and its context. But since many IoT applications also require timely response times, such semantic reasoning has to happen nearly in real-time, and hence reasoning has to be done over data/event streams, a.k.a Stream Reasoning. This work presents a semantic model and an IoT middleware service for data stream reasoning using a combination of Complex Event Processing (CEP) and continuous queries over streams of RDF semantic data. In our approach we use CEP in the IoT Edge devices to derive time-annotated RDF data from the basic events, and C-SPARQL for processing online queries over a domain ontology, aiming to check if some indirect fact or property about the system or the environment can be deduced. In order to showcase and test the capacity of nearly real-time inference over data streams of our middleware service, we designed and implemented a scenario of fire warning in a hypothetical multi-store building with temperature and humidity sensors.