CommonRoad-Prediction

A collection and interface for CommonRoad-based prediction algorithms.

Project status

Currently implemented and tested models:

• Constant Velocity Linear Predictor [1]
• Constant Velocity Curvilinear Predictor [1]
• Constant Acceleration Linear Predictor [1]
• Constant Acceleration Curvilinear Predictor [1]

In development:

• Intelligent Driver Model (IDM) Predictor [2]
• Lane-Changing Model MOBIL Predictor [3]

We highly welcome your contribution. If you want to contribute a prediction algorithm, please create an issue/pull request in our GitHub repository.

Installation and Usage

We recommend to use PyCharm (Professional) as IDE.

Usage in other projects

We provide an PyPI package which can be installed with the following command

pip install commonroad-prediction

Development

It is recommended to use poetry as an environment manager. Clone the repository and install it with poetry.

git clone git@github.com:commonroad/commonroad-prediction.git
poetry shell
poetry install

Examples

We recommend to use PyCharm (Professional) as IDE. An example script for visualizing predictions is provided here.

Documentation

You can generate the documentation within your activated Poetry environment using.

poetry shell
mkdocs build

The documentation will be located under site, where you can open index.html in your browser to view it. For updating the documentation you can also use the live preview:

poetry shell
mkdocs serve

Authors

Responsible: Roland Stolz, Sebastian Maierhofer

References

The implemented algorithms are based on the subsequent publications:

[1] R. Schubert, E. Richter and G. Wanielik, "Comparison and evaluation of advanced motion models for vehicle tracking," Proc. of the IEEE Int. Conf. on Information Fusion, 2008, pp. 1-6.

[2] M. Treiber, A. Hennecke, and D. Helbing, "Congested traffic states in empirical observations and microscopic simulations," Physical Review E, vol. 62, no. 2, pp. 1805–1824, 2000.

[3] A. Kesting, M. Treiber, and D. Helbing, “General lane-changing model MOBIL for car-following models,” Transportation Research Record, vol. 1999, pp. 86–94, Jan. 2007