Voxel512026cc-by-4.0
HapTile (FiftyOne multimodal MCAP)
A haptic-informed vision-tactile-language-action dataset of a UR5e robot with a Robotiq gripper performing contact-rich tabletop manipulation tasks under teleoperation, with synchronized RGB-D, tactile sensor, and haptic feedback data.
Downloads0
Episodes1699
Hours12.36
Why This Matters for Physical AI
HapTile provides rich multimodal data combining vision, tactile sensing, and haptic feedback for training robots to perform complex contact-rich manipulation tasks with human-level dexterity and force awareness.
Technical Profile
- Modalities
- rgbdepthtactileproprioceptionlanguagehaptics
- Robot Embodiments
- UR5eRobotiq 2F-85 gripper
- Action Space
- joint_positions
- Environment
- lab
- Task Types
- manipulationgraspingpick_and_placeinsertionstackingfoldingwipingpouringturning
- Episodes
- 1699
- Total Hours
- 12.36
- Data Format
- MCAP
- Annotation Types
- language_instructionshaptic_feedback_labelsaction_labels
- License
- cc-by-4.0
Access
Need custom rgb data?
Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.
Request a Sample Pack