DiffeRT2d 0.4.1

2D Toolbox for Differentiable Ray Tracing

DiffeRT2d

Differentiable Ray Tracing Python Framework for Radio Propagation.

DiffeRT2d is built on top of the JAX library to provide a program that is differentiable everywhere. With that, performing gradient-based optimization, or training Machine Learning models with Ray Tracing (RT) becomes straightforward! Moreover, the extensive use of the object-oriented paradigm facilitates the simulation of complex objects, such as metasurfaces, and the use of more advanced path tracing methods.

The objective of this tool is to provide a simple-to-use and highly interpretable RT framework for researchers engaged in fundamental studies of RT applied to radio propagation, or any researcher interested in the various paths radio waves can take in a given environment.

IMPORTANT: For 3D scenarios at city-scales, checkout DiffeRT.

• Installation
  • Dependencies
  • Pip install
  • Install From Repository
• Usage
• Citing
• Contributing
  • Reporting an Issue
  • Seeking for Help
  • Contact

Installation

While installing DiffeRT2d and its dependencies on your global Python is fine, we recommend using a virtual environment (e.g., venv) for a local installation.

Dependencies

DiffeRT2d uses JAX for automatic differentation, which in turn may use (or not) CUDA for GPU acceleration.

If needed, please refer to JAX's installation guidelines for more details.

Pip Install

The recommended way to install the latest release is to use pip:

pip install differt2d

Install From Repository

An alternative way to install DiffeRT2d is to clone the git repository, and install from there: read the contributing guide to know how.

Usage

For a quick introduction to DiffeRT2d, check you our Quickstart tutorial!

You may find a multitude of usage examples across the documentation or the examples folder, or directly in the examples gallery.

Citing

If you use this software, please cite it as:

@article{Eertmans_DiffeRT2d_A_Differentiable_2024,
  title   = {{DiffeRT2d: A Differentiable Ray Tracing Python Framework for Radio Propagation}},
  author  = {Eertmans, Jérome and Oestges, Claude and Jacques, Laurent},
  year    = 2024,
  month   = jun,
  journal = {Journal of Open Source Software},
  volume  = 9,
  number  = 98,
  pages   = 6915,
  doi     = {10.21105/joss.06915},
  url     = {https://joss.theoj.org/papers/10.21105/joss.06915}
}

or use the Cite this repository button on this page to get other citation formats.

Contributing

Contributions are more than welcome! Please read through our contributing section.

Reporting an Issue

If you think you found a bug, an error in the documentation, or wish there was some feature that is currently missing, we would love to hear from you!

The best way to reach us is via the GitHub issues. If your problem is not covered by an already existing (closed or open) issue, then we suggest you create a new issue.

The more precise you are in the description of your problem, the faster we will be able to help you!

Seeking for help

Sometimes, you may have a question about , not necessarily an issue.

There are two ways you can reach us for questions:

• via the GitHub issues;
• or via GitHub discussions.

Contact

Finally, if you do not have any GitHub account, or just wish to contact the author of DiffeRT2d, you can do so at: [e-mail].