Reproduce “Certifiably Safe Manipulation of Deformable Linear Objects via Joint Shape and Tension Prediction” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Yiting Zhang, Shichen Li

Researchers at the University of Michigan and University of Illinois developed a framework that integrates DLO shape and tension prediction with certifiable safety guarantees for manipulating deformable linear objects in contact-rich environments. This approach prevents DLO overextension and robot collisions, achieving a 76% task success rate in simulated wire harness assembly.

Key insights

  • Explicitly modeling and integrating DLO internal tension prediction is crucial for achieving certifiably safe manipulation, significantly reducing overextension risks compared to shape-only prediction methods.
  • Rigorous uncertainty propagation via polynomial zonotopes enables formal safety guarantees for both the robot and DLO, even when integrating learned predictive models with their inherent uncertainties.
  • A balanced approach that accounts for and manages necessary contact interactions, rather than overly avoiding them, is essential for robust and successful DLO manipulation in complex environments.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.