Hybrid Knowledge Modeling for Ambient Intelligence

摘要

This paper describes our research in enhance everyday devices as a solution to adapt Spoken Dialogue Systems (SDS) within ambient intelligence. In this context, a SDS enables universal access to ambient intelligence for anyone, anywhere at anytime, allowing the access to any device through any media or language. The main problem that we want to address is the spontaneous configuration of SDS to deal with a set of arbitrary plug and play devices. Such problem is resumed as a portability feature and is a critical research issue. We propose a hybrid approach to design ubiquitous domain models to allow the SDS to recognize on-the-fly the available devices and tasks they provide. When a device is activated or deactivated, a broker's knowledge model is updated from device's knowledge model using a knowledge integration process. This process was tested in the home environment represented by a set of devices.