Stochastic Modeling of Lightweight Floating Objects

We propose a stochastic modeling framework for light-weight objects floating in flows. Realistic and turbulent motions are conveniently provided for real-time animation or gaming environment, which is achieved by utilizing a stochastic differential equation (SDE) to model the momentum change of the objects. The model incorporates adaptive random fluctuation by adopting physical turbulence models computed from a pre-generated flow field. This base flow defines the major route of a light-weight object. Meanwhile, the jiggling behavior is implemented as a SDE-modeled random process on-the-fly, which enables very fast computation and easy control. These features are critical to integrate such natural phenomena into interactive 3D animation applications, which was previously hindered due to the conflict between limited computational resources and high-cost numerical simulation.