Using semantic technologies to design a Spatio-Temporal database

This research deals with the problem of the enormous amounts of data that is streaming from the ‘sensor web’ into our computer systems. This data is useless to us unless it is properly stored, queried and presented by means of computer systems. These are typically databases, query engines, applications and user interfaces. In particular the manner how data is stored in a database determines what information you can retrieve from the system. This is widely known when it comes to traditional relational databases, however when complex data structures such as spatial, temporal and spatio-temporal structures are involved the user of sensor data simply lacks an understanding of this issue. This thesis work addresses this problem by designing a prototype of an expert system that automatically makes the selection of an appropriate technical solution based on information entered by the user of sensor data. For the design of this prototype, techniques coming from Geo Information science are combined with those from Artificial Intelligence and Computer Science. The Artificial Intelligence techniques used are the building of an ontology and the use of a reasoning system and rule engine. These techniques, called semantic techniques, are typically useful for storing concepts and relations, querying them and drawing conclusions from them. These techniques are compared with one modelling method commonly used in software engineering, namely UML. The prototype that was built enables the user to enter requirements for the question to be posed upon the sensor data, as well as information about the structure of the data set used. The prototype also contains knowledge of existing technical solutions as well as features of space, time and space-time. A logical component then decides which technical solution is selected based on the information entered by the user. For the logical component two different semantic technologies were possible. One is the use of a reasoner. The other solution is a rule engine. Both solutions were implemented and this resulted in two prototypes, the JessTabDemo and the ReasonerDemo. The capabilities of the two prototypes were evaluated against the predefined prototype requirements. The conclusion was drawn that neither solution satisfied every predefined requirement. For reasons of comparison an imaginary solution based on UML was envisioned. This solution also did not satisfy every predefined requirement. This research concludes therefore that the envisioned instrument can best be built with a combination of UML and semantic technologies. It remains a challenge for the future to combine static solutions, such as databases, with dynamic ones such as ontologies.GIMAGIS TechnologyOTB Research Institute for the Built Environmen