Context-dependent car navigation as kind of human-machine collaborative interaction

This is a paper about one of our research area's most evasive concepts, namely that of “context”. In HCI as well as UbiComp, it has inspired many projects venturing to exploit the semantics of individual users' contexts. That is lead by intentions to study the potential of human-machine collaborative interaction and make applications behave more efficiently, user-friendly and invisibly. This has not considerably turned into a new paradigm of computing, we believe, due to a not completely satisfied theoretical model of what context really is. This paper intends to alleviate that by specifying a model of context that is even more personal, and by that better adapted to the particular domain of in-car navigation. The model is next used to implement an application, which demonstrates its usefulness. Future work shall take this experiment further, to support the inner-local context-aware functionality of such systems in practice.