IDEAS-Lab-Northwestern2026cc-by-4.0

ManiGuard Grasp Annotations

A curated database of 1,547 annotated 6-DoF grasps over 221 object instances for the Franka Panda gripper, used for automated demonstration generation across six ManiGuard-Bench task families.

Downloads0
Episodes1547

Why This Matters for Physical AI

This grasp annotation database enables systematic evaluation of robotic manipulation safety and provides diverse, manually-curated grasps for training and benchmarking manipulation systems.

Technical Profile

Robot Embodiments
Franka Panda
Action Space
end_effector_pose
Environment
simulation
Task Types
graspingmanipulation
Episodes
1547
Data Format
json
Annotation Types
grasp_annotationsapproach_hints
License
cc-by-4.0
Part of the ManiGuard family

Access

Need custom physical AI data?

Claru builds purpose-built datasets for simulation applications with dense human annotations and quality assurance.

Request a Sample Pack