Declan O’Sullivan, Dave Lewis, Vincent Wade

Enabling Adaptive Semantic Interoperability for Pervasive Computing

Trinity College Dublin (Ireland)

Embedding processors, sensors and actuators in everyday products implies an explosion in the number and type of organisations that need to be involved in achieving the seamless interoperability implied by the pervasive computing vision. Many of the network interoperability problems can be addressed by Internet protocols and XML data encoding. However, the potential for debilitating heterogeneity of semantics in application level interoperability remains. In this paper we examine the potential for using ontology-based semantics for supporting run-time semantic interoperability between pervasive computing applications. Automatic semantic interoperability between ontologies representing arbitary, separately developed models is beyond the abilites of current semantic reasoning techniques. Thus, we present a software and process framework (called OISIN) that guides the design time identification of candidate mapping information for a set of ontologies and enables runtime determination of the actual mappings based on the context of the interoperability required. We demonstrate its operation with a "visiting lecturer" university case study.