EvA - A Self Adaptable Event-Based Recognition Framework for Three-Dimensional Activity Zones

With recent advancements in supporting fields like Embedded Systems and Ambient Assisted Living (AAL), intelligent environments are becoming reality. Learning and adapting to user behaviors and gaining some basic knowledge about the underlying user intentions and activities are essential features of an intelligent system. Many systems use an underlying environment model with spatial and semantic information, and often the manual creation of such spatial semantic models is an error-prone and time-consuming task. Moreover, these approaches often neglect how people actually use their physical space. The concept of Activity Zones defines the environmental context by regions of similar user activities and can be learned by observing human behavior. In this paper, we present our approach to extend the notion of Activity Zones by applying a three-dimensional zone computation and visualization process that is independent from the present smart home infrastructure and that is able to autonomously adapt to changes in the environment as well as to shifts in the user behavior.