Self-supervised Physics-Informed Manipulation of Deformable Linear Objects with Non-negligible Dynamics
Self-supervised Physics-Informed Manipulation of Deformable Linear Objects with Non-negligible Dynamics
This paper introduces SPiD, a framework combining an extended mass-spring model with augmented self-supervised training to handle dynamic manipulation of deformable linear objects. The authors demonstrate the method on rope stabilization and trajectory tracking tasks in both simulation and real-world settings, showing robustness to distribution shifts, unseen initial states, varying rope properties, and external disturbances using a markerless perception approach.
The authors present SPiD, a physics-informed self-supervised learning framework for manipulating deformable linear objects like ropes, demonstrating sim-to-real generalization on rope stabilization and trajectory tracking tasks.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
Source：arXiv Robotics — research abstracts · arxiv.org