Qualitative Spatial Query Processing : Towards Cognitive Geographic Information Systems

For a long time, Geographic Information Systems (GISs) have been used by GIS-experts to perform numerous tasks including way finding, mapping, and querying geo-spatial databases. The advancement of Web 2.0 technologies and the development of mobile-based device applications present an excellent opportunity to allow the public -non-expert users- to access information of GISs. However, the interfaces of GISs were mainly designed and developed based on quantitative values of spatial databases to serve GIS-experts, whereas non-expert users usually prefer a qualitative approach to interacting with GISs. For example, humans typically resort to expressions such as the building is near a riverbank or there is a restaurant inside a park which qualitatively locate the spatial entity with respect to another. In other words, the users' interaction with current GISs is still not intuitive and not efficient. This dissertation thusly aims at enabling users to intuitively and efficiently search spatial databases of GISs by means of qualitative relations or terms such as left, north of, or inside. We use these qualitative relations to formalise so-called Qualitative Spatial Queries (QSQs). Aside from existing topological models, we integrate distance and directional qualitative models into Spatial Data-Base Management Systems (SDBMSs) to allow the qualitative and intuitive formalism of queries in GISs. Furthermore, we abstract binary Qualitative Spatial Relations (QSRs) covering the aforementioned aspects of space from the database objects. We store the abstracted QSRs in a Qualitative Spatial Layer (QSL) that we extend into current SDBMSs to avoid the additional cost of the abstraction process when dealing with every single query. Nevertheless, abstracting the QSRs of QSL results in a high space complexity in terms of qualitative representations