Configurable semi-autonomic animated animal characters in interactive virtual reality applications

Abstract

In many virtual reality (VR) simulation and training applications, it is desirable and even critical to have computer controlled animal characters that can behave with a high degree of realism. The realism can be measured in two aspects. One is behavioral realism, or how real the characters act and respond to the commands and environments. The other aspect is visual realism. Many efforts have been directed at animating human characters as well as other animal characters in applications such as interactive computer games. However, more research is still needed for realistic animation of animal characters due to the vast variety of animal species and different application purposes. In some of the environments for which security personnel are being trained, animals, such as dogs, are part of the training programs. So we need to include realistic behavior and visual representations of animal characters in our training applications. This study is focused on animal behavior and animation for VR applications. A simple implementation of a real-time animal animation method that is configurable is proposed to make it easy for user interaction. Synthetic animal characters are included in the system so that their behavior can be programmed and controlled digitally. A two-stage state machine is used in the system. The top-level state machine controls the animal behavior. The secondary state machine controls animation. Animation blending and procedural control are used to make the animation smooth.