Ropedia Releases Xperience 2: 1,500+ Scanned Real-World Scenes Paired With Egocentric Demos
liuziwei7 · x · 2026-11-16
Ropedia unveiled Xperience 2, billed as a gold-standard data foundation for scaling physical AI, pairing egocentric human demonstrations with metric-scale 3D reconstructions of the spaces they were recorded in.
Key details:
- Scale: 1,500+ scanned 3D scenes across 596 locations (homes, stores, restaurants, car parks, streets), 12 scene categories, 4,000+ egocentric recordings, 74 tasks, 131 hours of first-person video
- Capture pipeline: a handheld LiDAR scanner captures each space in one pass at metric scale; operators then wear HOMIE Gen 2 (head-mounted, 4 fisheye cameras, IMU, 4-channel audio) to record multiple tasks in the same scene
- Core value: long after recording, robots can be placed in a simulation of the same space and trained on tasks never demonstrated
- Quality control: every reconstruction is human-reviewed and graded (perfect / usable / reuse candidate)
- Benchmarks: evaluated across depth & geometry (11 models), 3D object detection (7 methods), and spatial intelligence (12 models, 500 questions); Boxer topped all three 3D detection metrics
Dataset, paper, and code are marked "coming soon".
Related event: Ropedia Launches Xperience 2, a Benchmark Dataset for Physical AI(3 posts)→
More from Embodied
- Ant Lingbo Embodied AI Challenge crowns TARS team after 48-hour real-robot hackathon — 新智元 · 2026-11-16
- Tesla FSD fleet tops 7 billion city miles, adding ~500M in under four weeks — XFreeze · 2026-11-16
- Hangzhou's Manycore Tech launches Aholo MS07 3D laser scanner with CHCNAV — teortaxesTex · 2026-11-16
- LeRobot 0.7 adds YAM, reBot B601 and Unitree G1 with NVIDIA SONIC support — mishig25 · 2026-11-16
- MathWorks open-sources a GitHub repo of robotics resources for arms, rovers and drones — lukas_m_ziegler · 2026-11-16
- Open-source CUDA-accelerated 3D LiDAR localization with 3x faster point cloud accumulation — 4310sy · 2026-11-16