ACE-Data-0: 150-Hour Multimodal Dataset for Embodied AI
Yukang Cao · hf · 2026-07-31
To address the data bottleneck in embodied intelligence, researchers introduce ACE-Data-0, a dataset designed to capture the full perception-action loop, including egocentric views, whole-body motion, dexterous manipulation, and multisensory signals over time.
Data Engine & Collection:
- Uses the Ambient Capture Engine (ACE) to transform real home environments into spatially calibrated, synchronized recording studios.
- Operates at table-scale (hand-object manipulation) and room-scale (whole-body locomotion).
- Records egocentric/exocentric video, full-body and hand motion, object 6-DoF trajectories, audio, and tactile signals as a unified stream.
Scale & Evaluation:
- Comprises 150 hours and 17M frames across 200 task categories, with 75,000 interaction episodes from 50 participants.
- Covers atomic manipulations, long-horizon household activities, and human-scene interactions.
- Introduces a hierarchical benchmark revealing that state-of-the-art methods struggle significantly with contact, occlusion, and long temporal horizons.
More from Embodied
- Stone Tao Joins Physical Intelligence to Focus on Robotics Simulation — svlevine · 2026-07-31
- VLA safety evaluation doc released, omits architecture and training — YuXiang_IRVL · 2026-07-31
- IROS 2026 Origami Challenge: Largest Visuotactile Robot Dataset Released — chris_j_paxton · 2026-07-31
- Thor: Open-Source 3D Printed Robotic Arm Under €350 with ROS2 Support — tom_doerr · 2026-07-31
- Sunday Robotics demos fully autonomous robot with real-time error correction — saranormous · 2026-07-31
- Unitree's Aug 10 IPO: How the Market Prices China's Next Productivity Leap — yongqianme · 2026-07-31