Geospatial Reasoning: Basic Concepts and Theory

Geospacial Reasoning’ is spacial reasoning including geographic data. This is an extremely board area. In this deliverable we presented our approaches to cover enough aspects from this area such that it is useful for Semantic Web applications. An important building block is a hierarchy of graphs which combines very low level coordinate based computations with abstract symbolic reasoning. Another building block is the ontology of transport networks OTN. The graphs and the corresponding algorithms, together with OTN, provide the built-ins for the specification language MPLL. MPLL is a functional programming language, which, however, is mainly used to control the application of built-in algorithms, for example shortest path computations. With MPLL one can define customized locational notions, and evaluate them over the given data (maps, transport networks, user context etc.). The processing of uncertainty is certainly a major issue also for locational reasoning. So far we have investigated the use of fuzzy sets in this context. The FuTIRe library, which was developed for fuzzy temporal intervals can be used for one-dimensional fuzzy distributions in a higher dimensional space as well. It is easy to integrate the processing of two-dimensional fuzzy distributions which are only represented by their membership functions. This is sufficient as a first approximation to fuzzy reasoning in two dimensions. Keyword List semantic web, geospatial notions, geospatial reasoning techniques Project co-funded by the European Commission and the Swiss Federal Office for Education and Science within the Sixth Framework Programme.