Robotics: Science and Systems XVIII

Rapid Locomotion via Reinforcement Learning

Gabriel B. Margolis*, Ge Yang*, Kartik Paigwar, Tao Chen, Pulkit Agrawal
* These authors contributed equally

Abstract:

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 leveraged from prior work. Videos of the robot’s behaviors are available at https://agility.csail.mit.edu/.

Download:

Bibtex:

  
@INPROCEEDINGS{Margolis-RSS-22, 
    AUTHOR    = {{Gabriel B.} Margolis AND Ge Yang AND Kartik Paigwar AND Tao Chen AND Pulkit Agrawal}, 
    TITLE     = {{Rapid Locomotion via Reinforcement Learning}}, 
    BOOKTITLE = {Proceedings of Robotics: Science and Systems}, 
    YEAR      = {2022}, 
    ADDRESS   = {New York City, NY, USA}, 
    MONTH     = {June}, 
    DOI       = {10.15607/RSS.2022.XVIII.022} 
}