Ubiquitous Computing - Computing in Context

Possible sensing technologies, in particular low level physical sensors, and perception techniques are assessed and their value for providing context in Ubiquitous Computing systems is analysed. Abstractions on sensor level, cue level, and context level are introduced, resulting in a flexible context acquisition architecture. A bottom-up approach for modelling context aware systems is introduced. This makes use of the fact that context or domain knowledge is more general on the level of artefacts, than on the system level. The creation of context aware systems, based on this approach, is then investigated using the method of prototyping. To generalise and communicate results, a pattern language for context aware systems is suggested. As context acquisition systems are mostly specific to a certain task, building such systems involves designing and building hardware and software. The research presented here shows methods, architectures, and tools to make the development process more efficient. The Smart-Its platform, a rapid prototyping system for context-aware Ubiquitous Computing systems, is introduced and use experience is reported. The observation that context naturally surrounds us, led to the development of a communication platform. This platform provides an effective means to distribute and receive information based on spatial and temporal relationships of components. In this research the notion of implicit human computer interaction, and in particular the use of context information as implicit input, is introduced. The implications on the user interface and on the human computer interaction process are analysed, as context is fundamentally different from events in user interfaces. Finally the research presents an overview on how Ubiquitous Computing systems can be evaluated. Different techniques are assessed, and the concept of probing users and developers with prototypes is presented.