An Automatic Generation Method of Student Behaviors in a Virtual Space

Abstract

In generation of human behaviors in a virtual space, an ordinal method is to generate human behaviors according to predefined action patterns, however, in this case it requires a lot of work for preparing action patterns. Therefore, we have been researching an automatic generation method of student behaviors, which respond to the environment change. This paper describes an improvement of user interface by reducing parameters necessary for defining of behaviors, and by using a library specified for person actions. In addition, when it is expected for two persons to cross on their way, this method makes one person to avoid the other with a natural and short route. As a result, the proposed method enabled us to generate various kinds of student behavior in a virtual campus.