Autonomous perception and behavior control for virtual human

Abstract

Nowadays virtual humans are being increasing used in different domains and have attracted more and more research attentions. The focus of virtual human research has recently shifted from geometric modeling and animation generation towards autonomous behavior and spontaneously interaction in virtual environment. This paper proposes an integrated framework for the autonomous perception and behavior control of virtual human, which mainly implements an Octree frustum culling-based visual perception system and a decision network-based behavioral decision making system. Experimental results in simulated environment show that the proposed approach is sufficiently robust and flexible for modeling autonomous virtual human.