CAPGrasp: An R3xSO(2)-equivariant Continuous Approach-Constrained Generative Grasp Sampler

Zehang Weng*, Haofei Lu, Jens Lundell, Danica Kragic
Division of Robotics, Perception and Learning (RPL), KTH
IEEE Robotics and Automation Letters (RA-L)
Paper Video Code

Our method learns to generate approach-constrained grasp poses for parallel-jaw grippers based on 3D partial point clouds. We conducted our study in both simulation and real world.

Abstract

We propose CAPGrasp, an equivariant 6-DoF continuous approach-constrained generative grasp sampler. It includes a novel learning strategy for training CAPGrasp that eliminates the need to curate massive conditionally labeled datasets and a constrained grasp refinement technique that improves grasp poses while respecting the grasp approach directional constraints. The experimental results demonstrate that CAPGrasp is more than three times as sample efficient as unconstrained grasp samplers while achieving up to 38% grasp success rate improvement. CAPGrasp also achieves 4-10% higher grasp success rates than constrained but noncontinuous grasp samplers. Overall, CAPGrasp is a sample-efficient solution when grasps must originate from specific directions, such as grasping in confined spaces.

Video

Approach-constrained grasping from a table

Approach-constrained grasping from a shelf

BibTeX

@article{weng2023capgrasp,
  author    = {Weng, Zehang and Lu, Haofei and Lundell, Jens and Kragic, Danica},
  title     = {{CAPGrasp}: An R3xSO(2)-equivariant Continuous Approach-Constrained Generative Grasp Sampler},
  journal   = {arXiv preprint arXiv:2310.12113},
  year      = {2023}
}