Differentiable Fluids with Solid Coupling for Learning and Control

Authors

  • Tetsuya Takahashi Adobe University of Maryland at College Park
  • Junbang Liang University of Maryland at College Park
  • Yi-Ling Qiao University of Maryland at College Park
  • Ming C. Lin University of Maryland at College Park

DOI:

https://doi.org/10.1609/aaai.v35i7.16764

Keywords:

Behavior Learning & Control

Abstract

We introduce an efficient differentiable fluid simulator that can be integrated with deep neural networks as a part of layers for learning dynamics and solving control problems. It offers the capability to handle one-way coupling of fluids with rigid objects using a variational principle that naturally enforces necessary boundary conditions at the fluid-solid interface with sub-grid details. This simulator utilizes the adjoint method to efficiently compute the gradient for multiple time steps of fluid simulation with user defined objective functions. We demonstrate the effectiveness of our method for solving inverse and control problems on fluids with one-way coupled solids. Our method outperforms the previous gradient computations, state-of-the-art derivative-free optimization, and model-free reinforcement learning techniques by at least one order of magnitude.
AAAI-21 Proceedings Cover

Downloads

Published

2021-05-18

How to Cite

Takahashi, T., Liang, J., Qiao, Y.-L., & Lin, M. C. (2021). Differentiable Fluids with Solid Coupling for Learning and Control. Proceedings of the AAAI Conference on Artificial Intelligence, 35(7), 6138-6146. https://doi.org/10.1609/aaai.v35i7.16764

Issue

Vol. 35 No. 7: AAAI-21 Technical Tracks 7

Section

AAAI Technical Track on Intelligent Robots