Virtual Humanoids: Let Them be Autonomous without Losing Control
Angela Caicedo, Daniël Thalmann · Infoscience (Ecole Polytechnique Fédérale de Lausanne) · 2000
There have been several efforts to build life-like autonomous creatures in virtual worlds, but only few of them have focused their intentions in presenting human-like autonomous creatures. In this paper we discuss the problem of building autonomous virtual humanoids with goal directed behaviors. We present the decision-making as a process compound of: goal achievement planning, dynamic belief management, evolving goals, internal states and confidence levels. As simulating real humans, we applied to our model theories of Humans' Trust to be able to interact with the virtual humanoids and direct them at real time. Keywords. Agent-oriented approach, goal achievement, dynamic beliefs, decision-making, planning, trust, confidence. 1 Introduction Virtual environments are becoming more important every day especially in educational and entertainment programs. With this evolution, the user gets more demanding, and the realism of this world is a key point. Users want to control and in...