A software architecture for ontology-driven situation awareness

Abstract

Human operators of large-scale control systems face the problem of information overload induced by the large amount of information provided by multiple heterogeneous and highly-dynamic information sources. Situation-aware information systems support operators by the aggregation of the available information to meaningful situations. Ontologies are a promising technology for realizing such systems, because of their semantically-rich kind of knowledge representation. The cross-cutting role of ontologies and the streaming character of situation awareness, however, challenge the design of an appropriate software architecture. In this paper, we propose a domain-independent software architecture based on a core ontology for situation awareness which leverages the reusability and the scalability of involved software components. This is achieved by the application of the well-known software architecture pattern pipes-and-filters. The proposed architecture is demonstrated by examples from the field of road traffic management. In addition, we contribute several lessons learned which should be helpful for developing ontology-driven information systems in general.