Distributed Episode Control System for Interactive Narrative Entertainment

摘要

We propose the massive action control system (MACS) for interactive narrative entertainment. MACS determines the action priorities for characters based in part on their own internal states, such as the motivation behind the action, feeling, and personality. MACS selects a behavior control module, called an episode tree, of about 1000 events, which is divided into action types based on these internal states and external situations. We demonstrate the effectiveness of the system with the Spilant World interactive animation contents at the National Museum of Emerging Science and Innovation in Japan, and NAMCO amusement park.