RoboMoCap releases 200+ hour egocentric vision × 120Hz mocap dataset for humanoid robot learning
DJiafei · x · 2026-11-09
RoboMoCap, a joint release from GroundedSI, BitRobot Network and NOKOV, is the largest locomotion-centric egocentric vision × optical motion capture dataset for humanoid robot learning.
- Scale: 200+ hours of recordings, 14 participants, 290+ diverse actions
- Sensing: six head-mounted and two wrist-mounted cameras (plus IMU, magnetometer, WAV audio), a 53-marker NOKOV optical mocap setup tracking whole-body motion at 120 Hz with 10mm accuracy — capturing leg and foot movement outside camera views — plus object 3D scans and hand pose
- Differentiation: most egocentric datasets focus on hand–object interactions; RoboMoCap centers walking, turning, obstacle navigation, carrying, and loco-manipulation
The dataset is publicly available (Hugging Face), the paper is out, code is coming soon, and the team is presenting at CoRL 2026.
Related event: RoboMoCap: 200+ hours of motion capture data for humanoid robots(3 posts)→
More from Embodied
- Machenta demos instant robot learning at CoRL 2026, founded by Imperial College one-shot learning team — Ed__Johns · 2026-11-10
- NVIDIA Unveils GR00T 2 Open-World Action Model, Two Papers Shortlisted for CoRL Best Paper — danfei_xu · 2026-11-10
- Nomagic Team Headed to CoRL 2026 to Dig Into Physical AI Mastery — m_wulfmeier · 2026-11-10
- ROBIB Meetup #3 in Ghent to Debate Build-vs-Buy for Humanoid Robots — MarwaEldiwiny · 2026-11-10
- CoRL 2026 kicks off in Austin, Texas, running November 9-12 — shaohua0116 · 2026-11-10
- Apple's HALO humanoid trains in 1000 simulated homes, hits 91% success in real houses — heyshrutimishra · 2026-11-10