Action planning for virtual human performances

Abstract

We describe an action planning approach that integrates a system of autonomous dialogue engines with narration and presentation modules to create a multi-modal storytelling environment. The goal is to let multiple virtual and human participants act on equal footing in a story that allows interactions that are flexible, but also constrained to ensure the success of the narrative goals. The model allows to combine scripted and planned behavior. The state of the story goals is continuously reflected to the narration module to allow it a fine-grained control over the development of the story. © Springer-Verlag Berlin Heidelberg 2005.