Event
RoboMIND benchmark releases multi-embodiment manipulation data including Tiangong humanoid trajectories
Key points
- RoboMIND introduces a multi-embodiment manipulation dataset with 107,000 demonstration trajectories across 479 tasks and 96 object classes.
- The Hugging Face dataset card specifies 19,152 trajectories from the Tien Kung humanoid robot, alongside Franka, AgileX, and UR-5e embodiments.
- It matters because RoboMIND is a major first-party dataset and benchmark tied to the Beijing center's Tiangong/Tien Kung humanoid platform.
Company context
National-level humanoid robotics innovation center focused on developing general-purpose humanoid systems through coordinated research, shared infrastructure, and industrial collaboration.
Context
- Segment
- Humanoid
- Event type
- Research Publication
- Geography
- Beijing · China