PUB - Publikationen an der Universität Bielefeld

In many smart environments including smart homes, human machine interactions are crucial. Being able to describe, store and query interaction data is an important feature. This especially gains importance when artificial embodied agents (i.e. robotic companions) are present. In order to support querying interaction data at a conceptual level, abstracting from specific and heterogeneous data schemas, we present an ontology for the domain of interaction. To model the interaction domain, the ontology imports two pre-existing ontologies: the Semantic Sensor Network Ontology and the Timeline Ontology. The ontology further includes interaction-relevant concepts as they occur in smart homes but also in other embodied interactive smart environments. We also describe a reference application for a data management and query system for interaction data that builds on this ontology. Our goal is to support the easy querying of interaction concepts both by machine agents but also by developers of applications running in smart environments. We therefore further provide exemplary queries to show the retrieval capabilities of a system that uses our ontology as a proof-of- concept