EUVIMEDEUVIMED entdecken
Crossref

Rapid locomotion via reinforcement learning

Gabriel B. Margolis, Ge Yang, Kartik Paigwar, Tao Chen, Pulkit Agrawal

The International Journal of Robotics Research · 2024-04 · 43 (4) · 572-587

DOI: 10.1177/02783649231224053 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The International Journal of Robotics Research, online erschienen am 02.01.2024. Verfasst von Gabriel B. Margolis und 4 weiteren Autor:innen. Die Arbeit wurde bisher 145-mal zitiert (Crossref) und nennt 49 Literaturangaben. Gefördert von: MIT-IBM Watson AI Lab, United States Air Force Research Laboratory, National Science Foundation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

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/ .

Förderung

  • MIT-IBM Watson AI Lab
  • United States Air Force Research Laboratory
  • National Science Foundation
  • Defense Advanced Research Projects Agency
  • DARPA Machine Common Sense Program
  • United States Air Force Artificial Intelligence Accelerator

Zitieren

APA

Margolis, G. B., Yang, G., Paigwar, K., Chen, T., & Agrawal, P. (2024). Rapid locomotion via reinforcement learning. The International Journal of Robotics Research, 43(4), 572-587. https://doi.org/10.1177/02783649231224053

Vancouver

Margolis GB, Yang G, Paigwar K, Chen T, Agrawal P. Rapid locomotion via reinforcement learning. The International Journal of Robotics Research. 2024;43(4):572-587. doi:10.1177/02783649231224053

BibTeX
@article{Margolis2024,
  author = {Margolis, Gabriel B. and Yang, Ge and Paigwar, Kartik and Chen, Tao and Agrawal, Pulkit},
  title = {Rapid locomotion via reinforcement learning},
  journal = {The International Journal of Robotics Research},
  year = {2024},
  volume = {43},
  number = {4},
  pages = {572--587},
  doi = {10.1177/02783649231224053}
}
Kostenlos starten →