Reasoning About Liquids via Closed-Loop Simulation
By Connor Schenck, Dieter Fox. 2017
29 citations recorded by the source corpus.
Simulators are powerful tools for reasoning about a robot's interactions with its environment. However, when simulations diverge from reality, that reasoning becomes less useful. In this paper, we show how to close the loop between liquid simulation and real-time perception. We use observations of liquids to correct errors when tracking the liquid's state in a simulator. Our results show that closed-loop simulation is an effective way to prevent large divergence between the simulated and real liquid states. As a direct consequence of this, our method can enable reasoning about liquids that would otherwise be infeasible due to large divergences, such as reasoning about occluded liquid.
Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Vision and Pattern Recognition
- Sergey Levine
- Luc Van Gool
- Li Fei-Fei
- Kristen Grauman
- Trevor Darrell
- Dacheng Tao
- Jitendra Malik
- Xiaogang Wang
In Control and Systems Engineering
- Sergey Levine
- Nicolas Manfred Otto Heess
- Josh Merel
- Nando de Freitas
- Yuke Zhu
- Ziyun Wang
- Saran Tunyasuvunakool
- Raia Hadsell
Citations in the corpus
- Latent Space Physics: Towards Learning the Temporal Evolution of Fluid Flow (2018, background): “…architecture for two-dimensional rigid bodies physics [CUTT16], while others have targeted predictions of liquid motions for robotic con-trol [SF17] Farimani et al. [FGP17] proposed an adversarial training approach to infer solutions for two-dimensional physics problems, such as heat…”
- Reinforcement and Imitation Learning for Diverse Visuomotor Skills (2018, background): “Modeling and reasoning about deformable objects and fluids is a long-standing challenge in the robotics community [41].”
- Generating Liquid Simulations with Deformation-aware Neural Networks (2017, background): “Learning physical functions is an area of growing interest within the research community, with applications ranging from physical priors for computer vision problems [21], over robotic control [36], to fast approximations for numerical solvers [39].”