EUVIMEDEUVIMED entdecken
Crossref

Heavy metal acrobatics: Learning dynamic control for skid-steer loaders

Pierre-Jean Lorentz iD, Pascal Egli iD, Edo Jelavic iD, Cliff Li, Ramitha Sundar iD, Harikrishnan Suresh, Paul Montgomery, Lee Redden, Matthias Müller iD, Ozan Sener iD, Vladlen Koltun iD, Marco Hutter iD

Science Robotics · 11.11.2026 · 11 (120)

DOI: 10.1126/scirobotics.aef0894 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Science Robotics, erschienen am 11.11.2026. Verfasst von Pierre-Jean Lorentz und 11 weiteren Autor:innen. Beteiligte Einrichtung: Robotic Systems Lab (RSL), ETH Zürich, Zürich 8092, Switzerland. Die Arbeit nennt 39 Literaturangaben. Gefördert von: Swiss National Science Foundation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Automation of heavy machinery offers the opportunity to improve productivity in mechanized industries such as construction, mining, and material handling. Yet, automation of such machines is challenged by their complex dynamics, in part because of hydraulic actuation that involves fluids, flexible elements, and nonlinear delays. A concomitant challenge is the size and weight of these machines, which can make experiments on the physical systems exceedingly dangerous. Here, we show that reinforcement learning (RL) in simulation, followed by transfer of the learned policies from simulation to reality, can successfully automate acrobatic maneuvers that push a heavy construction machine to the limits of its actuation and have only previously been accomplished by a small number of elite human operators in official world-record performances. We trained policies that perform autonomous acrobatic maneuvers on a John Deere 332G, an operational 5-tonne skid-steer loader. The policies can autonomously swing the machine up onto two wheels and, while balancing on two wheels, drive over rough terrain, trace out specified trajectories, spin around, and swing the bucket back and forth before landing it safely back onto four wheels. This work demonstrates that RL and sim-to-real transfer enable dynamic autonomous control of heavy machinery, directly contributing to greater automation in heavy equipment sectors.

Autor:innen und Institutionen

  • Robotic Systems Lab (RSL), ETH Zürich, Zürich 8092, Switzerland.
  • Blue River Technology (BRT), Santa Clara, CA 95054, USA.
  • Gravis Robotics AG, Opfikon 8152, Switzerland.
  • Intelligent Systems Lab, Intel, Jackson, WY 83001, USA.

ORCID-Profile: Pierre-Jean Lorentz, Pascal Egli, Edo Jelavic, Ramitha Sundar, Matthias Müller, Ozan Sener, Vladlen Koltun, Marco Hutter

Förderung

  • Swiss National Science Foundation

Zitieren

APA

Lorentz, P. J., Egli, P., Jelavic, E., Li, C., Sundar, R., Suresh, H., Montgomery, P., Redden, L., Müller, M., Sener, O., Koltun, V., & Hutter, M. (2026). Heavy metal acrobatics: Learning dynamic control for skid-steer loaders. Science Robotics, 11(120). https://doi.org/10.1126/scirobotics.aef0894

Vancouver

Lorentz PJ, Egli P, Jelavic E, Li C, Sundar R, Suresh H, et al. Heavy metal acrobatics: Learning dynamic control for skid-steer loaders. Science Robotics. 2026;11(120). doi:10.1126/scirobotics.aef0894

BibTeX
@article{Lorentz2026,
  author = {Lorentz, Pierre-Jean and Egli, Pascal and Jelavic, Edo and Li, Cliff and Sundar, Ramitha and Suresh, Harikrishnan and Montgomery, Paul and Redden, Lee and Müller, Matthias and Sener, Ozan and Koltun, Vladlen and Hutter, Marco},
  title = {Heavy metal acrobatics: Learning dynamic control for skid-steer loaders},
  journal = {Science Robotics},
  year = {2026},
  volume = {11},
  number = {120},
  doi = {10.1126/scirobotics.aef0894}
}
Kostenlos starten →