Interactive environments and the interaction with artificial characters can be traced back to the first days of computing with the introduction of computer games. However, interaction, virtual worlds and animation have gone far beyond these simplistic programs to develop sophisticated understanding and models of synthetic behaviour, dynamics and high-level non-deterministic control. ALIVE [36], a graphical virtual reality environment integrates complex behavioural mechanisms and vision input to synthesize a large space of emergent behaviours. Additionally, computer graphics research in kinematics, physical simulations and dynamics [71] have produced interesting and compelling synthesis and interactions with virtual humans and animals. This area includes work by Badler [3], Thalmann [17] and Terzopolous [58].
The synthesis and acquisition of behaviour has emerged in robotics as an important design paradigm and deserves mention. Evolving beyond the first order design of dynamics and control systems, robotics also utilizes artificial intelligence techniques including adaptivity, path planning, potential functions and navigation models. However, building on more biological inspiration, more complex behaviours including human-like and life-like idiosyncrasies are being implemented [12] [38]. As well, multi-agent interaction and cooperation models have addressed higher order issues and involve more sophisticated behavioural modeling [40] [39].
Of course, the more physically compelling the synthesis, the more believable the more believable the synthetic characters. However, there is a caveat: good graphics and good robotics do not necessarily imply good behaviours, believable interactions or true behavioural learning.