PhysicalAI-Robotics-NuRec

Dataset Description The Physical AI NuRec dataset seeks to empower robotic researchers to build the next generation of physical AI based end-to-end robotic models. This dataset includes various 3DGUT in USD files that can be loaded in Isaac Sim. Some datasets also include a mesh and occupancy map. The Mesh components are used for collision detection while the 3DGUT components provide realistic rendering. The asset can also be used with Isaac Sim Extensions like MobilityGen for AMR data generation. Dataset Owner(s) Di Zeng, Chirag Majithia, Harel Omer, Sameer Chavan, Isaac Deutsch, Weihan Wang, Yuchen Deng Release Note 06/06/2025 for Nova Carter datasets 06/02/2025 for Zurich lounge dataset 10/07/2025 for Zurich fourth floor iphone dataset 10/29/2025 for Stereo hand-held Endeavor-Andoria, Endeavor-Livingroom, and Endeavor-Wormhole datasets Releases and Versions Dataset assets are versioned with Git tags on Hugging Face. Use the tag that matches your Isaac Sim version: Tag Description Compatible Isaac Sim Version v0.0 First version of the assets 5.1.0 and above v0.1 Second version of the assets 6.0.0 and above To check out a release locally after cloning the repository: git fetch origin --tags git checkout v0.0 # or v0.1 On Hugging Face, select the same tag from the revision dropdown when browsing or downloading files. License/Terms of Use cc-by-4.0 Intended Usage Open sourcing for Isaac Sim use cases as part of NVIDIA physical AI dataset initiative. The dataset will be downloadable via Hugging Face. Dataset Characterization Data Collection Method Nova Carter Datasets The raw data is collected with Nova Carter and ISAAC ROS Nova Recorder, conducted through systematic planning in a controlled indoor environment with optimal lighting conditions. The setup consisted of: Four calibrated stereo cameras mounted on the Nova Carter robot platform Strategic camera placement (front, left, right, back) to achieve comprehensive 360-degree coverage Pinhole OpenCV camera model with