Omniverse Visualization

The visualization component of biologger-sim uses NVIDIA Omniverse USD Composer to render a high-fidelity digital twin of the tracked animal.

Contents:

• Installation & Setup
  • Windows Setup (Visualization)
  • Ubuntu Setup (Headless Simulation)
  • Hybrid Workflow (WSL + Windows)
• User Guide
  • Coordinate Systems & Conventions
  • Features Overview
  • Telemetry Display (HUD)
  • Tracking & Visualization Modes
  • Camera & Navigation
  • Time Control & Analysis (Instant Replay)
  • Configuration
  • Multi-Entity Configuration
  • ZMQ Protocol (MessagePack)
  • High-Performance Scaling
• Technical Architecture
  • Architectural Overview
  • Implementation Details
  • Payload Structure
  • Deployment Considerations
  • Verification
  • Slip Angle Diagnostics
  • GPU Acceleration (NVIDIA Warp)
  • Coordinate System Reference

Key Features

• Real-time telemetry streaming via ZeroMQ at 60+ Hz

• Automatic NED → USD coordinate conversion with validated slip angle diagnostics

• Flexible mesh orientation supporting arbitrary asset authoring conventions

• Instant replay with full trajectory visualization and time-travel scrubbing

Architecture

The system uses a “Hybrid” architecture:

1. Simulation (WSL/Linux): Runs the Python pipeline and publishes telemetry via ZeroMQ.

2. Visualization (Windows/Omniverse): Subscribes to the ZeroMQ stream and updates the 3D scene.

Quick Reference

See also

Coordinate Systems & Conventions

Understanding NED vs USD coordinate systems and automatic conversion.

Mesh Orientation & Transform Stack

How animal mesh assets are aligned with USD conventions via spawn rotations.

Slip Angle Diagnostics

Diagnostic metrics for validating heading/velocity alignment.