For running the PI Racing matlab code you need the following libraries installed on your system:
- CMake
- Version >= 1.65 (
sudo apt install cmake
)
- Version >= 1.65 (
- Boost (for GTSAM)
- Version >= 3.0 (
sudo apt install libboost-all-dev
)
- Version >= 3.0 (
- GTSAM (see installation hint below)
- PI Racing Factors (from this repo)
The code was written and tested on Ubuntu 18.04.6 LTS and Matlab R2021b.
Clone the GTSAM repo from Github and change to its root directory, then:
git checkout wrap-export
mkdir build
cd build
cmake ..
sudo make install
Clone this project and change to its root directory, then:
cd src/cpp
mkdir build && cd build
cmake ..
sudo make install
- There is an example script given called pi_min_curv.m (minimum curvature path)
- Start the script by opening the src directory in Matlab and running the pi_min_curv.m script to see the optimization for the Berlin 2018 ePrix circuit
- Change the track_file variable to "modena_2019.csv" to see the results with comparison for the Modena racetrack
- The script expects the GTSAM matlab library to be in the directory /usr/local/gtsam_toolbox/
@article{pi-racing,
author = {Salman Bari and
Ahmad Schoha Haidari and
Dirk Wollherr},
title = {A Fast Approach to Minimum Curvature Raceline Planning via Probabilistic
Inference},
journal = {CoRR},
volume = {abs/2203.03224},
year = {2022},
url = {https://arxiv.org/abs/2203.03224},
eprinttype = {arXiv},
eprint = {2203.03224},
timestamp = {Thu, 10 Mar 2022 14:39:36 +0100}
}