Edinburgh Napier University

  In a pervasive environment, it is essential for computing applications to be context-aware. However, one of the major challenges is the establishment of a generic and dynamic context model. Many different approaches to modeling the context exist, but an application- and domain-agnostic context model, that captures various types of context information and the dependencies between them, that could be reused and shared by different applications, and that can be dynamically changed when a shift in focus occurs, is missing. Therefore, we are interested in defining a structure for the dynamic management of context information. This paper describes our notion of context and proposes a distributed context management architecture that supports the development of context-aware applications. It presents CANDEL, a generic context information representation framework that considers the context as a dynamic product line composed of context primitives (CPs). Frame based software product line techniques are used together with OWL ontology to define CPs and to dynamically generate the current context model. Further, using Petri-Nets, we also show how this framework will be used to support context-aware adaptive pervasive applications.