Using behavioral simulation to animate complex processes

A form of forward simulation termed "behavioral simulation" is presented as a solution to a restricted class of motion control problems in computer animation in which the domain to be animated can be represented as a large collection of one or more homogeneous sets of discrete interacting structures. The complexity of the movement of the system as a whole emerges as a result of simple local interactions between the discrete structures. Traditional approaches to motion control are first discussed to support the proposal that simulation is a useful means to procedurally generate motion in computer animated scenes. The use of reductionism as a means to deal with complexity in those scientific fields which are relevant to computer animation are reviewed in order to provide a background in which the technique of behavioral simulation can be presented, These relevant areas include physically-based modeling, cognitive modeling of computer animated characters, and actor-based approaches to software development in computer science. Behavioral simulation is then defined and the rationale and justification for using behavioral simulation is described. This technique is compared to other approaches that have been used to procedurally generate animated motion. The design for a test-bed which has been constructed for producing behavioral simulations is discussed as well as several motion models which have been implemented within this test-bed. The underlying models for each motion model are described. The advantages and pitfalls of using simulation in animation are illuminated by the results of experiments using these models. Descriptions of several animations that were actually produced using this simulation technique are presented as well as suggestions on how this technique can be integrated with traditional computer animation techniques.