Reproduce “Active Contact Sensing for Robust Robot-to-Human Object Handover” 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
Linfeng Li, Lin Shao, David Hsu
Researchers developed an active contact sensing framework that allows robots to robustly decide when to release an object during human handover. The method employs a probabilistic piecewise-linear model of human-object contact and small probing motions, achieving a 97.5% handover success rate, which is significantly higher than passive sensing baselines.
Key insights
- Proactively perturbing the human-object interaction provides richer, disambiguating information compared to merely observing, enabling robust differentiation between firm grasps and incidental contact.
- A probabilistic piecewise-linear contact model effectively captures the unilateral and non-smooth nature of physical contact, quantifying uncertainty for improved decision-making.
- The Bayesian framework enables principled assessment of grasp firmness and guides adaptive, information-gathering motions, optimizing for uncertainty reduction.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.