Plan Recognition in Instrumented Environments

Michael Schneider

In: Vic Callaghan, Achilles Kameas, Angélica Reyes, Dolors Royo, Michael Weber (editor). Proceedings of the 5th International Conference on Intelligent Environments. International Conference on Intelligent Environments (IE-09) July 20-21 Barcelona Spain Pages 295-302 Ambient Intelligence and Smart Environments 2 ISBN 978-1-60750-034-6 IOS Press 2009.


Plan recognition is the process of inferring a user's plans and goals by observing her behavior. With their sensors, instrumented environments provide a good infrastructure for the application of plan recognition, and in return can benefit from plan recognition knowledge by better supporting their users. This article addresses specific issues that arise if plan recognition is applied in physical environments. A plan recognition model based on dynamic Bayesian networks is presented and practically applied in an instrumented kitchen scenario. Furthermore, a utility function for sensor information is derived from this model, which can be used to solve the problem of sensor selection in real-world plan recognition applications.


German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz