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.