Implementing Ubiquitous Services with Ontologies: Methodology and Case Study

Modern Ubiquitous Information Systems (UISs) appear more and more as intelligent systems that provide business services for mobile users by taking into account both sensed data and administrative records stored on different servers. Currently, data integration is achieved by a centralized relational database where all the information coming from the remote servers is stored. To improve such architecture we have proposed to convert the relational archive into an RDF triple store where data are represented by standard terms interrelated by subject-predicate-object relations (also called data ontology). In this way, ubiquitous applications could be developed independently on the technology used to collect the data and on how the data are formatted on the various specialized servers. The paper proposes to implement such an architecture in a distributed environment to achieve higher reliability and time performance. Also, how structuring the service interface according to task ontology is outlined since this improves usability. Implementation issues illustrate how the software environments based on the paradigm Models-View-Controllers, e.g., Ruby on Rails powered by JQMobile and Flash Builder, may facilitate the implementation of the proposed methodology.