analiza aiqbench
This paper combines pretrained reinforcement learning policies with model-based generative trajectory optimization for contact-rich robot manipulation. It is a research preprint evaluated on three tasks (surface scanning, peg insertion, humanoid pushing), so its claims rest on those experiments and on how well the approach transfers to contact conditions outside training. For ML readers, the useful question is whether reusing RL policies as proposal generators, rather than as direct controllers, actually improves robustness under distribution shift.
Manifold Generative Annealing for Contact Constraint-aware Impedance Control with Reinforcement Learning Prior
arXiv:2610.21140v1 Announce Type: new Abstract: Contact-rich robotic tasks require coordinated motion and impedance under uncertain interaction dynamics and physical constraints. While recent methods on learned policies provide useful interaction strategies, they may fail when contact conditions differ from those encountered during training. Model-based generative methods show promising receding-horizon trajectories optimization results, but often lack structured behavioral guidance and refinement that jointly accounts for constraint geometry and actual contact responses. To this end, we propose Manifold Generative Annealing (MGA), a joint model-free and model-based generative framework for motion--impedance optimization. MGA first organizes pretrained reinforcement learning policies into structured sequence proposals, retaining task-specific interaction strategies without committing to their execution. It then evaluates these proposals alongside exploratory samples through model-based rollouts for refinement or separate execution selection. Crucially, realization-aware manifold refinement aligns updates from the refinement bank with both task constraints and locally achievable physical responses, allowing learned experience to guide online adaptation under changing contact conditions. Experiments cover surface scanning, peg insertion, and humanoid pushing, with improved safety success rate of impedance control under nominal and sensing-shift conditions.
Abstract:Contact-rich robotic tasks require coordinated motion and impedance under uncertain interaction dynamics and physical constraints. While recent methods on learned policies provide useful interaction strategies, they may fail when contact conditions differ from those encountered during training. Model-based generative methods show promising receding-horizon trajectories optimization results, but often lack structured behavioral guidance and refinement that jointly accounts for constraint geometry and actual contact responses. To this end, we propose Manifold Generative Annealing (MGA), a joint model-free and model-based generative framework for motion--impedance optimization. MGA first organizes pretrained reinforcement learning policies into structured sequence proposals, retaining task-specific interaction strategies without committing to their execution. It then evaluates these proposals alongside exploratory samples through model-based rollouts for refinement or separate execution selection. Crucially, realization-aware manifold refinement aligns updates from the refinement bank with both task constraints and locally achievable physical responses, allowing learned experience to guide online adaptation under changing contact conditions. Experiments cover surface scanning, peg insertion, and humanoid pushing, with improved safety success rate of impedance control under nominal and sensing-shift conditions.
Robotics (cs.RO)
arXiv:2610.21140 [cs.RO]
(or arXiv:2610.21140v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.21140 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 265 cuvinte
Subiecte conexe
Constraint-Decomposed Impedance Control from Generative Policy for Articulated Object Manipulation
arXiv · 14 oct. 2026
Digital States 2026: AI Moves From Experiment to Infrastructure
govtech.com · 24 sept. 2026
FMOPF: Latent Flow Matching with Constraint-Aware Interaction Priors for AC Optimal Power Flow
arXiv · 28 iul. 2026
PAKT: Physically-Aligned Kinesthetic Teaching for Reinforcement Learning
arXiv · 24 sept. 2026