The use of simulation techniques for motion control of rigid bodies in computer graphics

There is a need for a simulation based motion control system for computer animation. This need arises from the desire to give the animator a higher level of control of objects in the scene through a small set of parameters as well as to allow him to generate realistic motions of objects easily and effectively. Attaching real world properties to objects in addition to shape and light reflection allows the animator to generate motions of the objects by applying a set of equations that describe how the objects should move. This approach has a natural parallel in the way computer graphics has approached the problem of display algorithms, which is exemplified by the attempt to simulate the physics of light interacting with the virtual objects.; There are several reasons why simulation systems have not been used extensively in computer animation. First, the simulation techniques have not been sophisticated enough to model some of the complex motion problems encountered in animation. Where there are adequate techniques, the cost of doing the necessary numerical calculations have been large. Finally, the traditional simulation techniques have not given the animators adequate control over the motions generated. This dissertation addresses each of these problems in a simulation system for generating the motions of rigid bodies.; The first problem is addressed by modeling the complex interactions between objects (including continuous contacts) in a completely general and novel way using collision detection and impact dynamics.; The second problem is addressed in a number of ways. The rotational dynamics are simplified by solving the Euler equations in the principal axis reference frame. The external dynamics of kinematically driven articulated figures are solved principally by using conservation of momenta. The use of conservation of momenta in the dynamics analysis allows the solution for the motion involving only the velocities and not the accelerations.; The third problem is addressed by allowing the animator to control the motion in three ways: traditional use of initial parameters, kinematic scripts, and cue scripts. Kinematic scripts allow the user to control objects in the scene while the object affects (and is affected by) the environment. This allows integration of traditional kinematic techniques with dynamic simulation techniques.