[SIGGRAPH 2022 Talk] DiffCloth: Differentiable Cloth Simulation with Dry Frictional Contact

Описание: DiffCloth: Differentiable Cloth Simulation with Dry Frictional Contact. ACM Transactions on Graphics.
🌍: https://people.csail.mit.edu/liyifei/...
📃: https://people.csail.mit.edu/liyifei/...

Abstract: Cloth simulation has wide applications in computer animation, garment design, and robot-assisted dressing. This work presents a differentiable cloth simulator whose additional gradient information facilitates cloth-related applications. Our differentiable simulator extends a state-of-the-art cloth simulator based on Projective Dynamics (PD) and with dry frictional contact. We draw inspiration from previous work to propose a fast and novel method for deriving gradients in PD-based cloth simulation with dry frictional contact. Furthermore, we conduct a comprehensive analysis and evaluation of the usefulness of gradients in contact-rich cloth simulation. Finally, we demonstrate the efficacy of our simulator in a number of downstream applications, including system identification, trajectory optimization for assisted dressing, closed-loop control, inverse design, and real-to-sim transfer. We observe a substantial speedup obtained from using our gradient information in solving most of these applications.