HuG: Learning Humanoid Whole-Body Affordance-Aware Grasping with Active Perception
HuG: Learning Humanoid Whole-Body Affordance-Aware Grasping with Active Perception
This research presents HuG, a framework for whole-body affordance-aware grasping with active perception for humanoids. The system uses a Perceive-Align-Grasp pipeline to handle occlusions and maintain stability. By curating training data with controller-in-the-loop dynamics simulation, it rejects 73% of kinematically valid but dynamically unstable grasps. Trained entirely in simulation with zero real demonstrations, HuG achieves an 87.5% success rate on a multi-level shelf task in the real world and generalizes to novel objects and heights below its training distribution.
The paper introduces HuG, a simulation-trained framework for humanoid grasping that integrates active perception and kinodynamic filtering to achieve 87.5% real-world success on a multi-level shelf task without using real demonstrations.
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.
- Environment
- SimulationOpen source S4
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
g data with controller-in-the-loop dynamics simulation, rejecting 73% of kinematically valid grasps that would destabilize the deployment controller, so the policy learns kinodynamically feasible affordances. Trained entirely in simulation with zero real demonstrations, HuG attains 87.5% real-world success on a multi-level shelf task, generalizes to novel objects, and holds at least 73% success at every tested height
Open source S4
. Beyond generalizing within its training distribution, the policy extrapolates spatially: at scale in simulation it succeeds at table heights well below any seen during training, indicating that kinodynamically filtered synthetic data grounds affordances in object geometry and whole-body feasibility rather than in memorized configurations.
Open source S5
Source：arXiv Robotics — research abstracts · arxiv.org