Rapid locomotion via reinforcement learning
By Gabriel B. Margolis, Ge Yang, Kartik Paigwar, Tao Chen, Pulkit Agrawal. Robotics: Science and Systems Conference, 2022
382 citations recorded by the source corpus.
Agile maneuvers such as sprinting and high-speed turning in the wild are challenging for legged robots. We present an end-to-end learned controller that achieves record agility for the MIT Mini Cheetah, sustaining speeds up to 3.9 m/s. This system runs and turns fast on natural terrains like grass, ice, and gravel and responds robustly to disturbances. Our controller is a neural network trained in simulation via reinforcement learning and transferred to the real world. The two key components are (i) an adaptive curriculum on velocity commands and (ii) an online system identification strategy for sim-to-real transfer. Videos of the robot’s behaviors are available at https://agility.csail.mit.edu/.
Subfields: Artificial Intelligence, Biomedical Engineering
View Paper: https://www.semanticscholar.org/paper/b7cbf03974511d7855b4c01e3005f3e38bd1a136
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Zipeng Fu
- Christopher G. Atkeson
- Chelsea Finn
- Gabriel B. Margolis
- Jianren Wang
- Sergey Levine
- Pieter Abbeel
- J. Urgen Schmidhuber
In Biomedical Engineering
- Gabriel B. Margolis
- Ziwen Zhuang
- Soeren Schwertfeger
- Hang Zhao
- Xuxin Cheng
- Ge Yang
- Xiaolong Wang
- Jialong Li
Citations in the corpus
- Robot Parkour Learning (2023, background): “Recently, learning-based methods have shown robust performance on walking [8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 12, 21, 22, 23, 24, 25, 26, 27], climbing stairs [20, 28, 29, 30, 31, 32, 33], mimicking animals [34, 35, 36, 37, 38, 39] and legged mobile manipulation [40, 41, 42] by…”
- Walk These Ways: Tuning Robot Control for Generalization with Multiplicity of Behavior (2022, methodology): “It is necessary to train using auxiliary rewards that bias the robot to maintain a particular contact schedule, action smoothness, energy consumption, or foot clearance [1, 2, 3, 5, 6, 7], to compensate for the sim-to-real gap.”
- Open-TeleVision: Teleoperation with Immersive Active Visual Feedback (2024, positive): “Learning based methods have achieved great success in locomotion control with massive simulation data and Sim2Real transfer [36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48].”
- Extreme Parkour with Legged Robots (2023, neutral): “Some recent work demonstrates blind dynamic running and jumping using sim2real RL on quadrupeds [41] and bipeds [12], [13].”