Getting Started with Sionna RT

Tested on: Ubuntu 22.04, CUDA, RTX 4060, x86, Intel i9-13905H

This README guides you through:

1. Installing Sionna RT locally and running the official example
2. Creating and rendering your own custom scenes from OpenStreetMap and Blender

---

Part 1: Install Sionna RT Locally & Run the Example

Step 1: Install Sionna RT with Mitsuba Support

Make sure you have Python ≥ 3.8 and pip:

pip install sionna[rt]

This installs:

• Sionna RT
• Mitsuba 3 (scalar_rgb mode)
• All required dependencies for local ray tracing

Step 2: Run the Official Example Locally

Run

python sionna_example.py

You should get a series of plots in the output/ directory.

Official Colab Notebook

If you prefer to start in the cloud with zero setup, try the official Sionna RT introduction notebook:

Sionna RT Colab Tutorial

---

Part 2: Build Your Own Scene from OpenStreetMap

Step 1: Export from OpenStreetMap

1. Visit https://www.openstreetmap.org.
2. Navigate to the area you want to simulate.
3. Use the Export function to download the .osm file of the region.

---

Step 2: Install Blender 3.5

Install Blender via Snap, then update to version 3.6 LTS:

sudo snap install blender --classic
sudo snap refresh blender --channel=3.6lts/stable

---

Step 3: Import .osm and Configure Blender Add-ons

Follow this tutorial to import and convert OSM to a Mitsuba-compatible format:

YouTube Tutorial – Blender OSM + Mitsuba Export

Required Add-ons:

• Blender-OSM: For importing OpenStreetMap files.
• Mitsuba Blender (for Mitsuba 2 or 3): For exporting XML scenes.

Important: When exporting from Blender to Mitsuba format, set the correct up axis according to your scene configuration (Blender default is Z-up).

---

Step 4: Export to Mitsuba XML

After importing the .osm file and configuring the scene:

• Export the scene to an XML file compatible with Mitsuba.

---

Step 5: Load and Visualize with Sionna RT

Use the custom.py scprits whith the correct xml file and meshes or the Jupyter notebook sionna-rt.ipbyn to load your XML scene. Example code:

from sionna.rt import load_scene

scene = load_scene("your_scene.xml")
scene.preview()  # Launch the interactive preview GUI

Jupyter notebook alows you to get a interactive preview of the scene. Thanks to it it is easier to place the camera transmiter and reciever.

---

Step 6: Enable GUI in Jupyter Notebook

To use Sionna RT's interactive preview in a notebook:

sudo snap install jupyterlab-desktop --classic

Then use the sionna.ipbyn file. This will open an interactive viewer for inspecting your scene.

Eurecom

---

With this setup, you can build complex and realistic radio environments for simulation in Sionna RT starting from real-world map data.