🏎️ Curvature-Integrated MPCC Local Trajectory Planning

TL;DR: CiMPCC is a curvature-integrated MPCC-based local trajectory planner that reduces lap time in autonomous racing by explicitly accounting for track curvature.

This repository provides an implementation to deploy and visualize the trajectory planner proposed in the ITSC 2024 paper "Reduce Lap Time for Autonomous Racing with Curvature-Integrated MPCC Local Trajectory Planning Method". The main branch contains the F1tenth simulator and the CiMPCC planner.

Update:

The ROS environment for the current branch is Noetic under Ubuntu 20.04. For the version of melodic, please see the branch melodic_py27.

🪄 Quickstart

Two methods are provided to configure the runtime environment:

Run directly using the pre-built Docker image.
Reconfigure from scratch using Docker.

Start by cloning this repository to the host:

git clone https://github.com/zhouhengli/CiMPCC.git f1tenth_ws

🛠️ Configure

Either of the following two methods can be used to deploy the environment.

✅ Run directly using the pre-built Docker image

Alternatively, the Docker configuration can be pulled from Google Drive. Simply download it to your Linux system.

[1/2] Import the prebuilt_v1.0.tar file as a new image using Docker import:

docker import prebuilt_v1.0.tar prebuilt_v1.0

[2/2] Now, you can use the imported image to create and launch a new container:

docker run -it \
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  -e DISPLAY=$DISPLAY \
  -v /home/ddrx/f1tenth_ws:/home/ddrx/f1tenth_ws \
  -w /home/ddrx/f1tenth_ws \
  prebuilt_v1.0 \
  /bin/bash

✅ [Optional] Reconfigure from scratch using Docker

[1/3] Pull Docker image:

docker pull ros:noetic-robot-focal

[2/3] Set up a container:

docker run -it \
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  -e DISPLAY=$DISPLAY \
  -v <host_path>/f1tenth_ws:/home/ddrx/f1tenth_ws \
  ros:noetic-robot-focal \
  /bin/bash

[3/3] Set up the necessary dependencies in the corresponding container using the bash script:

cd /home/ddrx/f1tenth_ws/
chmod +x setup_env.sh
./setup_env.sh

🚀 Planning

[1/3] Set up a container and enter the following commands. After the final command, the map should pop up:

source /opt/ros/noetic/setup.bash && source /home/ddrx/f1tenth_ws/devel/setup.bash
./toolkit/sim_setup.sh -n map416

[Optional] If the RViz interface does not appear and there is an error qt.qpa.xcb: could not connect to display :0, it may be because Docker does not have access to the display server. Try the following command in the host machine: xhost +local:docker.

[2/3] Start a new container and run the CiMPCC planner:

docker exec -it <CONTAINER ID> /bin/bash
export PYTHONPATH=$PYTHONPATH:/home/ddrx/f1tenth_ws/toolkit/casadi/
source /opt/ros/noetic/setup.bash && source /home/ddrx/f1tenth_ws/devel/setup.bash
roslaunch nonlinear_mpc_casadi ddrx_sim_nmpcc.launch

[Optional] For a real vehicle, use the command: roslaunch nonlinear_mpc_casadi ddrx_nmpcc.launch.

[3/3] Use the '2D Nav Goal' as the starting signal for racing.

💻 Customization

This project allows for the customization of the map and track files used by the CiMPCC method, as well as the parameters. Adjustments can be made according to specific needs.

Before modifying them, replace the root_path in ./toolkit/params/*.yaml with the path to the track files.

✏️ Customized map and track

Maps files can be found in toolkit/maps/, and the map in this project is generated using Cartographer.

Track files are located in toolkit/tracks/, where path and boundaries are defined as a closed curve, and <track_name>_center_derivates.csv defines the deviations in the x and y directions.

✏️ Parameter Tuning

Parameter files are saved in toolkit/params/, where the parameter definitions are consistent with those described in the paper. Try tuning the parameters to improve the performance (reduce the lap time) of the planner 🏁.

Feel that parameter tuning is troublesome? Check out my latest work, which uses Bayesian optimization for automatic parameter tuning.

🤗 Acknowledgments

Many thanks to the excellent open-source repositories listed below:

Please contact Zhouheng Li if you have any questions or suggestions.

📑 Citations

If you find this project useful for your research, please consider citing the following paper and leaving a ⭐—both would be greatly appreciated :)

@INPROCEEDINGS{10920215,
  author={Li, Zhouheng and Xie, Lei and Hu, Cheng and Su, Hongye},
  booktitle={2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC)}, 
  title={Reduce Lap Time for Autonomous Racing with Curvature-Integrated MPCC Local Trajectory Planning Method}, 
  year={2024},
  pages={1066-1073},
  doi={10.1109/ITSC58415.2024.10920215}
}

CiMPCC

Install / Use

README