EUVIMEDEUVIMED entdecken
Crossref

RoboBallet: Planning for multirobot reaching with graph neural networks and reinforcement learning

Matthew Lai iD, Keegan Go iD, Zhibin Li iD, Torsten Kröger iD, Stefan Schaal iD, Kelsey Allen, Jonathan Scholz

Science Robotics · 03.09.2025 · 10 (106)

DOI: 10.1126/scirobotics.ads1204 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Science Robotics, erschienen am 03.09.2025. Verfasst von Matthew Lai und 6 weiteren Autor:innen. Beteiligte Einrichtung: Google DeepMind, London, UK. Die Arbeit wurde bisher 26-mal zitiert (Crossref) und nennt 34 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Modern robotic manufacturing requires collision-free coordination of multiple robots to complete numerous tasks in shared, obstacle-rich workspaces. Although individual tasks may be simple in isolation, automated joint task allocation, scheduling, and motion planning under spatiotemporal constraints remain computationally intractable for classical methods at real-world scales. Existing multiarm systems deployed in industry rely on human intuition and experience to design feasible trajectories manually in a labor-intensive process. To address this challenge, we propose a reinforcement learning (RL) framework to achieve automated task and motion planning, tested in an obstacle-rich environment with eight robots performing 40 reaching tasks in a shared workspace, where any robot can perform any task in any order. Our approach builds on a graph neural network (GNN) policy trained via RL on procedurally generated environments with diverse obstacle layouts, robot configurations, and task distributions. It uses a graph representation of scenes and a graph policy neural network trained through RL to generate trajectories of multiple robots, jointly solving the subproblems of task allocation, scheduling, and motion planning. Trained on large randomly generated task sets in simulation, our policy generalizes zero-shot to unseen settings with varying robot placements, obstacle geometries, and task poses. We further demonstrate that the high-speed capability of our solution enables its use in workcell layout optimization, improving solution times. The speed and scalability of our planner also open the door to capabilities such as fault-tolerant planning and online perception-based replanning, where rapid adaptation to dynamic task sets is required.

Autor:innen und Institutionen

  • Google DeepMind, London, UK.
  • University College London, London, UK.
  • Intrinsic, Mountain View, CA, USA.

ORCID-Profile: Matthew Lai, Keegan Go, Zhibin Li, Torsten Kröger, Stefan Schaal

Zitieren

APA

Lai, M., Go, K., Li, Z., Kröger, T., Schaal, S., Allen, K., & Scholz, J. (2025). RoboBallet: Planning for multirobot reaching with graph neural networks and reinforcement learning. Science Robotics, 10(106). https://doi.org/10.1126/scirobotics.ads1204

Vancouver

Lai M, Go K, Li Z, Kröger T, Schaal S, Allen K, et al. RoboBallet: Planning for multirobot reaching with graph neural networks and reinforcement learning. Science Robotics. 2025;10(106). doi:10.1126/scirobotics.ads1204

BibTeX
@article{Lai2025,
  author = {Lai, Matthew and Go, Keegan and Li, Zhibin and Kröger, Torsten and Schaal, Stefan and Allen, Kelsey and Scholz, Jonathan},
  title = {RoboBallet: Planning for multirobot reaching with graph neural networks and reinforcement learning},
  journal = {Science Robotics},
  year = {2025},
  volume = {10},
  number = {106},
  doi = {10.1126/scirobotics.ads1204}
}
Kostenlos starten →