Interpolation synthesis for articulated figure motion

Abstract

Realistic real time articulated figure motion is achieved by reprocessing a stored database of motions. Motions are created to exact specification by interpolation from a set of example motions, effectively forming a parameterized motion model. A pre-processing step involving iterative calculations is used to allow efficient direct computations at run time. An inverse kinematics capability is shown that is based on interpolation. This method preserves the underlying qualities of the data, such as dynamical realism of motion capture, while generating a continuous range of required motions. Relevant applications include networked virtual reality and interactive entertainment.