REAL-TIME AUTONOMOUS PEDESTRIAN NAVIGATION WITH GROUPS

摘要

Much research on modeling pedestrians has focused either on individual navigation and control or on simulating large crowds composed of many individuals. But many pedestrians travel in small groups of two to five or a few more. This paper describes improved real-time pedestrian behaviors that include support for small groups. We present our model for autonomous navigation of individuals, and then extend it to support small groups. We demonstrate our approach by using it to drive a commercial motion generation engine that produces high quality pedestrian animation.