Skip to content

platypus0121/Omniverse-Warehouse-Digital-Twin

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Omniverse Warehouse Digital Twin

A real-time warehouse digital twin built on NVIDIA Omniverse. Users can manage inventory (restock / remove cardboard boxes) directly in a 3D USD scene via a web-based dashboard, and query warehouse state using natural language powered by OpenAI.


Features

  • Real-time 3D visualization — warehouse scene streamed from Omniverse Kit to a web browser via WebRTC
  • Inventory control — restock or remove cardboard boxes on specific rows and shelves using a structured dashboard
  • USD scene manipulation — boxes are spawned as USD references and physically placed on shelf surfaces
  • AI natural language queries — ask questions about warehouse state in any language (e.g. "Which shelf is almost empty?") powered by OpenAI GPT-4o-mini
  • Shelf status variants — shelves automatically update visual state (empty / warning / normal / full) based on inventory level

Tech Stack

Category Technology
3D Engine NVIDIA Omniverse Kit
Scene Format OpenUSD (Universal Scene Description)
Streaming Omniverse WebRTC (omni.kit.livestream.webrtc)
Kit Events carb.eventdispatcher, omni.kit.app
AI OpenAI API (gpt-4o-mini)
Frontend Framework React 18 + TypeScript
Styling Tailwind CSS v4
Build Tool Vite
Kit UI omni.ui

Project Structure

warehouse-digital-twin/
├── frontend/
│   └── src/
│       ├── App.tsx                  # Main app, stream state management
│       ├── WarehouseDashboard.tsx   # Inventory control + AI query dashboard
│       ├── StreamOnlyWindow.tsx     # WebRTC stream + custom event handler
│       └── AppStream.tsx            # AppStreamer wrapper
└── source/extensions/
    └── demo.semantic_layout_generator/
        ├── config/extension.toml    # Extension config & dependencies
        └── demo/semantic_layout_generator/
            └── extension.py         # Kit extension: event handling, USD scene, OpenAI

🚧 Project Status: Work in Progress (WIP)

About

A real-time warehouse digital twin built on NVIDIA Omniverse. Users can manage inventory—such as restocking or removing cardboard boxes—directly within a 3D USD scene via a web-based dashboard, and query the warehouse state using natural language powered by OpenAI

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors