GRIP
收藏资源简介:
GRIP数据集是一个大规模、通用的抓取数据集,由加州大学洛杉矶分校的AIVC实验室创建。该数据集包含100K高质量抓取姿态,涵盖不同场景下的软性UMI夹爪和刚性LEAP手与刚性及软性物体的交互。数据集记录了丰富的物体形状、大小和材料,以及高保真摩擦接触下的夹爪和物体的变形和应力分布。这个数据集可以用于神经抓取生成和应力场预测等应用。
The GRIP dataset is a large-scale, general-purpose grasping dataset developed by the AIVC Laboratory at the University of California, Los Angeles (UCLA). It encompasses 100,000 high-quality grasping poses, covering interactions between the soft UMI gripper and rigid LEAP hand, as well as their contacts with both rigid and soft objects across diverse scenarios. The dataset records rich variations in object geometries, dimensions and material compositions, alongside the deformation and stress distributions of both grippers and objects under high-fidelity frictional contact conditions. This dataset enables applications such as neural grasping generation and stress field prediction.

- 1GRIP: A General Robotic Incremental Potential Contact Simulation Dataset for Unified Deformable-Rigid Coupled Grasping加州大学洛杉矶分校(AIVC实验室) · 2025年



