An Architecture for Context Prediction

Today’s information appliances appear very powerful, featuring on-device storage and processing power, communication technology and supporting many different applications. Context awareness is currently considered as one of the key issues for future device generations, with context prediction being the next step in research. The goal is not only to recognize the current context of an information appliance or its user, but also to predict the future context and thus enable the device to become proactive. In this paper, an approach to recognize and predict high level context information from low level sensor data is presented. Targeting a wide range of platforms, this approach has also been implemented in a software framework for on-line, un-supervised context prediction.