nvaszabo's picture
Fixed forklift-pallet composition intersection, updated screeshot in readme
946b3c0
metadata
license: cc-by-4.0
language:
  - en
tags:
  - 3D
  - simready
  - warehouse
  - NVIDIA
pretty_name: NVIDIA Physical AI SimReady Warehouse OpenUDS Dataset
size_categories:
  - n<1K
task_categories:
  - image-segmentation
  - robotics
download_size: 10132336442
dataset_size: 10132336442
configs:
  - config_name: default
    data_files: physical_ai_simready_warehouse_01.csv
    image_field: asset_preview

NVIDIA Physical AI SimReady Warehouse OpenUSD Dataset

Physical AI SimReady Warehouse Dataset Preview

Dataset Version: 1.0.0
Date: March 18, 2025
Author: NVIDIA, Corporation License: CC-BY-4.0 (Creative Commons Attribution 4.0 International)

Contents

This dataset includes the following:

  • This README file
  • A CSV catalog that enumerates all of the OpenUSD assets that are part of this dataset including a sub-folder of images that showcase each 3D asset (physical_ai_simready_warehouse_01.csv). The CSV file is organized in the following manner:
    • asset_preview : A visual preview (thumbnail) of the asset
    • asset_name : Name of the asset
    • relative_path: location within archive of the asset Usd file
    • classification: Asset Classification (Prop, Scenario, Assembly, etc.)
    • label: Label name of q-code in English.
    • q_code: Semantic Label (derived from WikiData Q-Code)
    • mass(kg): Approximate mass of the object in kilograms
    • thumbnail_path: path to an image thumbnail (PNG format)
  • 578 OpenUSD 3D assets that are compatible with NVIDIA's IsaacSim application for simulation. The dataset archive is approximately 10GB in size.

Prerequisites

This dataset is intended for use within NVIDIA IsaacSim version 4.x, and depicts an industrial warehouse environment.

For users who are new to OpenUSD as a format, NVIDIA encourages the following learning resources to get you started:

For users who are new to IsaacSim and simulation, and want to learn how to add robots to your scenes (including the scenario in this dataset), here are some additional learning resources to explore:

How to Use

Adding Rigid Body Physics to the Dataset Assets

To begin, open the physical_ai_simready_warehouse_01.usd file that is in the root of the dataset archive. This is a complete warehouse scene that has been pre-built as a starting point for simulation. This scene file uses OpenUSD composition arcs to payload in many other individual OpenUSD files.

  • NOTE: The initial load of this file might take several minutes as IsaacSim will need to compile all of the materials and physics colliders.

All of the 3D assets in this scene have USDPhysics Colliders applied, but no Rigid Bodies are applied. It is up to the user to add rigid body physics to individual assets as desired to satisfy their simulation goals. The application of a rigid body to an asset will determine if it should fall with gravity, collide with other assets and be interactive with other elements in the scene like a robot.

To add Rigid Bodies to an asset, do the following within IsaacSim.

  1. Select any asset you wish to apply Rigid Bodies to in the Stage Panel. Be sure to select the top-level payload XForm prim that represents the entire 3D asset, not one of the geometry or material prims within it.

    Physical AI Dataset Preview
  2. In the Properties Panel, disable the Instanceable checkbox to make the asset unique.

  3. Next, in the Properties Panel, click Add -> Physics -> Rigid Body to add the RBD physics to the asset.

    Physical AI Dataset Preview
  4. Repeat this process for each asset you want to interact with when a simulation is run.

  5. When the assets are configured with Rigid Body Dynamics, hit the Play button in IsaacSim to begin the simulation.

Asset Classifications

  • Prop: A single object that can be placed in a scene.
  • Assembly: A collection of props that can be placed in a scene as a logical group.
  • Scenario: A collection of props and assemblies to make an environment for simulation. (ex: A warehouse)

Additional Questions

If you have additional questions, you are encouraged to join our NVIDIA Omniverse Discord channel to interact with other Omniverse and IsaacSim users.