Situation-based testing for ubiquitous computing systems

Abstract

A common trend in modern applications is the move towards more mobile, adaptive, customisable software. The evolution of software from static, invariant tools for narrow portions of a task to adaptive, open interaction frameworks is embodied in the use of a variety of technologies for creating a reconfigurable application. The key challenge to improving application behaviour in response to external knowledge is in making the representation of that external knowledge compatible with the application’s representation. This external information, relevant to the user and their task is commonly known as context information. In the past, context information has typically been integrated using an a-priori model of context, which constrains the type of information which can be used as context. This thesis presents a model for context integration which does not depend on an a-priori model of context. This thesis presents a novel approach to integrating contextual information through the use of a context mediator based on ontology mediation. This context-informed semantic interoperation approach is based on the exchange of both schema and instance data, in the form of ontologies, between heterogeneous sources of context and a target application. The mediator represents the collective knowledge of a contextual situation by linking ontologies in their native form through a shared semantic view. The approach is innovative in that it combines user-defined and ontological reasoning to provide a more expressive method for bridging differences in representations between different sources and their target without an a-priori model. It demonstrates the use of semantic interoperation as an approach to allowing richer knowledge exchange between applications and their surroundings.