Robust high-resolution cloth using parallelism, history-based collisions, and accurate friction

Area of Science:

• Computer Graphics
• Computational Physics

Background:

• Simulating high-resolution cloth requires efficient handling of complex interactions.
• Existing models often struggle with accurate cloth-object and self-collision dynamics.

Purpose of the Study:

• To develop a more accurate and efficient method for simulating high-resolution cloth.
• To improve the fidelity of cloth simulations by enhancing collision and friction models.

Main Methods:

• Simulated cloth with up to 2 million triangles.
• Proposed an accurate cloth-object friction model.
• Developed a robust history-based repulsion/collision framework.
• Utilized distributed memory parallelism and addressed Gauss-Seidel ordering for collision response.

Main Results:

• Achieved highly detailed folds and wrinkles in cloth simulations.
• Demonstrated accurate and efficient handling of cloth-object and self-collisions.
• Successfully performed high-resolution and high-fidelity simulations.

Conclusions:

• The proposed methods significantly improve the realism and detail of cloth simulations.
• The framework is efficient and scalable for complex, high-resolution scenarios.
• This work advances the state-of-the-art in physically-based animation and simulation.