Tool Support for Probabilistic Intention Recognition Using Plan Synthesis

摘要

To provide assistance in intelligent environments it is necessary to accurately infer the users needs and wishes. In this demonstration we present a probabilistic plan recognition system that is able to track the user and to compare different hypotheses about the users behavior and her goal(s) based on observations of the current activity. Furthermore, the tool provides a probability distribution over the possible goals and selects the most probable hypothesis as the user intention.