commonroad-crime 0.4.5

criticality measures of automated vehicles

CommonRoad-CriMe

Toolbox to compute Criticality Measures (e.g. time-to-collision, time-to-react,...). Such measures can be used to trigger warnings and emergency maneuvers in driver assistance systems or repair an infeasible trajectory. If you have questions or want to report problems or suggestions, please start a Github discussion / Github issue.

🚀 Installation Guide

commonroad-crime can be installed with:

$ pip install commonroad-crime

Develop CriMe locally

For adding new measures, we recommend using Anaconda to manage your environment so that even if you mess something up, you can always have a safe and clean restart. A guide for managing python environments with Anaconda can be found here.

After installing Anaconda, create a new environment with:

$ conda create -n commonroad-py310 python=3.10 -y

Here the name of the environment is called commonroad-py310. You may also change this name as you wish. In such case, don't forget to change it in the following commands as well. Always activate this environment before you do anything related:

$ conda activate commonroad-py310
or
$ source activate commonroad-py310

Then, install the dependencies with:

$ cd <path-to-this-repo>
$ pip install -e .
$ conda develop .

To test the installition, run unittest:

$ cd tests
$ python -m unittest -v

To get started your journey with our criticality measures, check the tutorials and the following tips.

Add new criticality measure

1. create a new branch with feature-<measure-name> and checkout the branch
2. navigate to commonroad_crime/data_structure/type.py to find the correct category of the measure and add an enumeration entry <abbreviation>: <explanation>
3. navigate to commonroad_crime/measure to find the above-mentioned category and create a python file named <abbreviation>.py. Then create a class inheriting the CriMeBase under commonroad_crime/data_structure/base.py
4. similar to other measures, you need to implement the compute() and visualize() functions

Define configuration parameters of the measure

1. navigate to commonroad_crime/data_structure/configuation.py to find the above-mentioned category and add a new instance to the class as self.<parameter> = config_relevant.<parameter>
2. you can then directly call the values using self.configuration.<category>.<parameter> in your measure class
3. to override the default parameter values, create a yaml file (name it the same as the scenario) in ./config_files and modify the values there

🚧 Documentation

The documentation of our toolbox is available on our website: https://cps.pages.gitlab.lrz.de/commonroad/commonroad-criticality-measures/.

Build documentation locally

In order to generate the documentation via Sphinx locally, run the following commands in the root directory:

$ pip install -r ./docs/requirements_doc.txt
$ cd docs/sphinx
$ make html

The documentation can then be launched by browsing ./docs/sphinx/build/html/index.html/.

🌟 Contributors (in alphabetical order by last name)

• Liguo Chen
• Marius Erath
• Florian Lercher
• Yuanfei Lin
• Sebastian Maierhofer
• Ivana Peneva
• Kun Qian
• Oliver Specht
• Sicheng Wang
• Youran Wang
• Zekun Xing
• Ziqian Xu

🔖 Citation

If you use commonroad-crime for academic work, we highly encourage you to cite our paper:

@InProceedings{lin2023crime,
      title     = {{CommonRoad-CriMe}: {A} Toolbox for Criticality Measures of Autonomous Vehicles},
      author    = {Yuanfei Lin and Matthias Althoff},
      booktitle = {Proc. of the IEEE Intell. Veh. Symp.},     
      pages     = {1-8}, 
      year      = {2023},
}

If you use this project's code in industry, we'd love to hear from you as well; feel free to reach out to Yuanfei Lin directly.