RFUniverse: A Multiphysics Simulation Platform for Embodied AI
Abstract
Multiphysics phenomena, the coupling effects involving different aspects of physics laws, are pervasive in the real world and can often be encountered when performing everyday household tasks.
Intelligent agents which seek to assist or replace human laborers will need to learn to cope with such phenomena in household task settings. To equip the agents with such kind of abilities, the research community needs a simulation environment, which will have the capability to serve as the testbed for the training process of these intelligent agents, to have the ability to support multiphysics coupling effects.
Though many mature simulation software for multiphysics simulation have been adopted in industrial production, such techniques have not been applied to robot learning or embodied AI research. To bridge the gap, we propose a novel simulation environment named RFUniverse.
This simulator can not only compute rigid and multi-body dynamics, but also multiphysics coupling effects commonly observed in daily life, such as air-solid interaction, fluid-solid interaction, and heat transfer.
Because of the unique multiphysics capacities of this simulator, we can benchmark tasks that involve complex dynamics due to multiphysics coupling effects in a simulation environment before deploying to the real world. RFUniverse provides multiple interfaces to let the users interact with the virtual world in various ways, which is helpful and essential for learning, planning, and control. We benchmark three tasks with reinforcement learning, including food cutting, water pushing, and towel catching. We also evaluate butter pushing with a classic planning-control paradigm. This simulator offers an enhancement of physics simulation in terms of the computation of multiphysics coupling effects.
Appendix
Supplementary File for Reinforcement Learning Benchmark on Common tasks such as Navigation, Simple Manipulation and Multi-agent Collaboration. You may find the experiments in the Exp Section in this file.
Codebase: Unity Server Side Python Client Side
Forum: forum
Documentation: documentation
Bibtex
@INPROCEEDINGS{Fu-RSS-23,
AUTHOR = {Haoyuan Fu AND Wenqiang Xu AND Ruolin Ye AND Han Xue AND Zhenjun Yu AND Tutian Tang AND Yutong Li AND Wenxin Du AND Jieyi Zhang AND Cewu Lu},
TITLE = {{Demonstrating RFUniverse: A Multiphysics Simulation Platform for Embodied AI}},
BOOKTITLE = {Proceedings of Robotics: Science and Systems},
YEAR = {2023},
ADDRESS = {Daegu, Republic of Korea},
MONTH = {July},
DOI = {10.15607/RSS.2023.XIX.087}
}
Video Demo
Quantitative Comparison
Towel in the wind (Left: towel in RFU, Right: towel in Abaqus).
Physics configuration:Cloth Material: Nylon, Young's modulus:3GPa, Poisson Ratio:0.35, Density: 1140kg/m^3
Air: Density: 1.225 kg/m^3, Dynamic Viscosity: 1.81*10^-5 Pa*s, Wind flow speed: Given below.
2. Heat Transfer (Left: butter melting in RFU, Right: butter melting in Abaqus).
Physics configuration:
1. melting point: 333K
2. heating surface temperature: Given blow
3. thermal conductivity: solid - 0.2 W/(m*K), liquid - 0.17 W/(m*K)
4. density: solid - 950 kg/m^3, liquid - 920 kg/m^3
5. heat capacity at constant pressure: solid - 2200 J/(kg*K), liquid - 2600 J(kg*K)
6. dynamic viscosity: 10^-2 Pa*s
Franka Robotics Reinforcement Learning Results
Duplicate RL Baseline from Panda-Gym
Atomic Action Demos
Locomotion
Walk & Run
Climb
Navigation
Look Around
Rotate
Navigation
Wait
Teleport
Crouch
Manipulation
Single-hand
Grasp & Transport & Place
Open & Close
Cover & Uncover
Fold & Unfold
Push & Pull
Soak
Switch
Drop
Bi-manual
Grasp & Transport & Place
Open & Close
Cover & Uncover
Fold & Unfold
Drop
Stretch
Tool-based
Cut
Hammer
Plug in & Plug out
Flip
Scoop
Screw
Stir
Sweep
Wipe