C. Kleine-Cosack, T. Pl{\"o}tz and G. A. Fink
Proc. KNX Scientific Conference, 2008.
Recent projects in the ï¬eld of Ambient Intelligence (AmI) concentrate on technical aspects to integrate evolving ubicomp tech- niques into so called smart rooms and smart houses. The concept of context awareness is utilized to create what is called intelligence in such scenarios and the representation of user information on system level in general takes place within a general context model. While technologically impressive, a lack of portability to practical applications can be observed for the majority of such research projects. The tremendous dependency on special and expensive hardware limits the scope of application and prevents Ambient Intelligence for the masses. Additionally the technology centered view accompanied by an (implicit) data driven and non-active user model carries along diï¬erent usability concerns. As a result the user turns out to be a foreign object on system level, which is excluded from the systemâs communication layer and â ignoring the userâs skills â is merely handled as an obstacle in a technology centered system. In this paper we present a case study featuring a novel approach to the integration of humans into AmI environments to prepare the ground for a user centric â thus user friendly â infrastructure for smart devices. The key aspect of the concept which we call human centered AmI is a dynamic and active user model which creates a virtual doppelganger of the user on software level. Mapping the userâs capabilities and skills to system level, human services integrate seamlessly into the KNX enabled intelligent environment, allowing for transparent human computer interaction. Moreover, it also widens the scope of existing hardware and software components. In particular, the functionality of KNX devices in practical scenarios is complemented by human services and the presented approach. The concept of human centered AmI is put into eï¬ect within the perception-oriented intelligent environment FINCA. This smart conference room is based on standard hardware components and, on sensor level, focuses on visual and acoustic data. In consequence our concept is easily transferable between related scenarios putatively exhibiting diï¬erent hardware settings.