commonroad-clcs 2025.2.0

Curvilinear (Frenet) Coordinate System for motion planning.

CommonRoad Curvilinear Coordinate System (CLCS)

Curvilinear coordinate frames (also Frenet frames) are a widely used representation in motion planning for automated vehicles. Curvilinear frames are aligned with a reference path and Cartesian points (x, y) are described by coordinates (s, d), where s is the arc length and d the lateral deviation to the reference path. The coordinate transformation is performed via a projection on the reference path for which different methods can be used [1] [2].

This project software for constructing and using curvilinear coordinate frames for reference paths given as polylines. We offer:

• C++ backend for efficient coordinate transformations
• Python frontend for usage with CommonRoad scenarios
• various utility methods for processing reference path (e.g., smoothing, curvature reduction, resampling etc ...)

:wrench: Installation

We provide two installation options for CommonRoad-CLCS: Installation as a Python package (recommended) or building from source.

1. Python Package: Install the python package via pip in your Conda environment:

   pip install commonroad-clcs
   

2. Build from source: To build the project from source and install it in your Conda environment, please refer to the README_FOR_DEVS. This option is only recommended for advanced users who would like to use or modify the C++ backend directly.

:link: System Requirements

The software is written in Python 3.10 and C++17, and was tested on Ubuntu 20.04 and 22.04. It should be compatible with later versions. For building the code, the following minimum versions are required:

• GCC and G++: version 10 or above
• CMake: version 3.28 or above.
• Pip: version 24.2 or above

We recommend using Anaconda as a package manager for Python.

:rocket: Getting Started

To get started, please see the minimal example in ./tutorials/clcs_minimal_example.py.

For more details, please refer to the Jupyter notebook tutorials provided in ./tutorials/.

:books: Documentation

A documentation including APIs is available on our website at: https://cps.pages.gitlab.lrz.de/commonroad/commonroad-clcs/.

To build the documentation locally, please install the optional docs listed in pyproject.toml first. The documentation can then be built using mkdocs via:

mkdocs build

You can browse the doc by launching ./docs/site/index.html

:busts_in_silhouette: Authors

Contributors (alphabetic order by last name): Peter Kocsis, Edmond Irani Liu, Stefanie Manzinger, Tobias Markus, Evald Nexhipi, Vitaliy Rusinov, Daniel Tar, Gerald Würsching

:speech_balloon: References

If you use our software for research, please cite:

@inproceedings{wursching2024robust,
    author={W{\"u}rsching, Gerald and Althoff, Matthias},
    title={Robust and Efficient Curvilinear Coordinate Transformation with Guaranteed Map Coverage for Motion Planning},
    booktitle={Proc. of the  IEEE Intelligent Vehicles Symposium},
    year={2024}
}

Additional references:

[1] Héry, Elwan, Stefano Masi, Philippe Xu, and Philippe Bonnifait. "Map-based curvilinear coordinates for autonomous vehicles." ITSC, 2017

[2] Bender, Philipp, Julius Ziegler, and Christoph Stiller. "Lanelets: Efficient map representation for autonomous driving." IV, 2014