Modelling of more realistic intelligent virtual agent in virtual and mixed reality

Abstract

. This article presents different options for modelling and realization of realistic intelligent virtual agents (IVA) in virtual and mixed reality. In relation to that a natural language dialogue with the users has been implemented; information about the preferences of the users is obtained, as well as their emotions, goals and the way, in which they prefer to achieve their goals. Emphasis is placed on the method for modelling the visualization of IVA; the facial expressions and emotions, which can be conveyed; the gaze of the eyes; the mouth and lip-sync with the spoken words; the gait and gestures. Some results of conducted user dialogues when solving a particular problem; the selection of a goal and a way of achieving it are discussed.