An Open Distributed Framework for Adaptive User Interaction in Ambient Intelligence

Challenges of handling user interaction in Ambient Intelligence environments are manifold. The systems installed in these environments are highly distributed with dynamic configurations in terms of integrated devices and installed applications. Context-awareness, personalization, and multimodality are critical for supporting more natural interaction and optimizing the interaction in an adaptive way. Research activities have dealt with different specific problems in the field and now it is high time for moving towards an open framework with a more comprehensive solution. This paper introduces results of such work with a high level of freedom in developing and deploying applications without needing to care about the available I/O infrastructure. The other way around, the latter can be changed without worrying about the application side. The independence of applications from the available I/O infrastructure helps to share mechanisms and manage such a complex scene more adequately. The key idea behind this framework is the natural distribution of tasks according to the real scene using a middleware solution supporting seamless connectivity and goal-based interoperability.