Grasp-Taxonomy Conditioning as a Mode Switch for Dexterous Grasping of Climbing Holds
Grasp-Taxonomy Conditioning as a Mode Switch for Dexterous Grasping of Climbing Holds
This paper introduces BetaGrasp, a 3D diffusion policy that uses a four-class grasp label (crimp, sloper, pinch, jug) to condition a DP3-style point-cloud policy on a Franka arm with a 16-DoF LEAP hand. The authors also present a climbing-hold benchmark using a spring testbed with ratchet slip-force readings. Analysis shows the point-cloud encoder alone fails to distinguish hold types, leading to incorrect jug grasps in simulation without the label; BetaGrasp achieves 94% correct grasp type versus 25% for the baseline. On 80 hardware trials, BetaGrasp increases average slip force from 9.6 N to 14.3 N and reduces fully slipped grasps from 44% to 10% of trials.
The study presents BetaGrasp, a 3D diffusion policy conditioned on grasp labels to improve dexterous manipulation of climbing holds on a Franka arm with a 16-DoF LEAP hand, reporting an increase in average slip force from 9.6 N to 14.3 N and a reduction in full slips from 44% to 10% across 80 hardware trials.
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.
Reported numbers
average slip force
9.6
View original evidence
cy performs. On 80 paired hardware trials, BetaGrasp raises average slip force from 9.6 N to 14.3 N, and grasps slipping fully off the hold drop from 44% of trials to 10%.
Open source S4average slip force
14.3
View original evidence
cy performs. On 80 paired hardware trials, BetaGrasp raises average slip force from 9.6 N to 14.3 N, and grasps slipping fully off the hold drop from 44% of trials to 10%.
Open source S4hardware trials
80
Reported trials
View original evidence
cy performs. On 80 paired hardware trials, BetaGrasp raises average slip force from 9.6 N to 14.3 N, and grasps slipping fully off the hold drop from 44% of trials to 10%.
Open source S4
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
arXiv:2611.12017v1 Announce Type: new Abstract: Dexterous grasping in robotics relies on distinct hand configurations depending on object geometry. We introduce BetaGrasp, a 3D diffusion policy that conditions a DP3-style point-cloud policy on a four-class grasp label (crimp, sloper, pinch, and jug), running on a Franka arm with a 16-DoF LEAP hand. We also introduce a climbing-hold benchmark: a spring testbed whose r
Open source S2
cy performs. On 80 paired hardware trials, BetaGrasp raises average slip force from 9.6 N to 14.3 N, and grasps slipping fully off the hold drop from 44% of trials to 10%.
Open source S4
atchet gives a slip-force reading in newtons. In our analysis the policy's point-cloud encoder did not separate the four holds, so the scan alone gave it little to choose a grasp from. Without the label (a DP3-style baseline), simulated rollouts from the robot's reset pose all end in a jug grasp, whatever the hold (25% correct grasp type, versus 94% for BetaGrasp). The label works as a switch for which grasp the poli
Open source S3
Source：arXiv Robotics — research abstracts · arxiv.org