ubm-rrt

This is an implementation of the Theta*-RRT algorithm for our RoboRacer (formerly F1Tenth) stack. The algorithm has been replicated from this paper:

L. Palmieri, S. Koenig and K. O. Arras

RRT-Based Nonholonomic Motion Planning Using Any-Angle Path Biasing

Since the package is not maintained anymore, documentation will not be updated according to UEP 11.

Authors

• UniBo Motorsport Driverless team
• @Scheggetta - for the overall package and RRT logic
• @TorioCrema - for the Theta* path generation algorithm and the speed profile generation
• @Noce99 - for the ROI generation from the map
• @SamueleCrimi - for the POSQ controller adaptation and implementation
• @andreaalboni - for the POSQ controller adaptation and implementation
• @ParsaMK - for the speed profile generation

How to make this ROS2 package work

First of all, this package has been tested on ROS 2 Foxy, but it should work without any problems also with the more recent versions like Humble and Jazzy.

Note

Only for organization members: These files can be directly copied into ubm-f1tenth repo to make the package work as of 12/22/2024. To automatically build it in the docker container, you would have to add the link in the ~/f1tenth_ws/src folder as usual.

Missing f1tenth_msgs package

The only package that is not available online that you need to build this package is indeed f1tenth_msgs. Though, just by understanding the code you can create your own messages' implementations.

What are map_filepath and map_yamlpath variables?

They are simply the filepaths of the map occupancy grid in .pgm format and its associated .yaml file, respectively. Those can be generated using the slam_toolbox repo.

What are ax_max_machines_path and ggv_path variables?

They correspond to the ax_max_machines.csv and ggv.csv file locations (filepaths), respectively. Those can be obtained and edited from the TUM racing line generation repo.

Build C++ dependencies

This package has some C++ dependencies that are exclusive to the rrt package: GEOS, FLANN, nanoflann and Armadillo. Once you have installed those, you can build the package as any other ROS 2 package using colcon build.

GEOS 3.12.1

cd ~ && \
wget https://github.com/libgeos/geos/releases/download/3.12.1/geos-3.12.1.tar.bz2 && \
tar xvfj geos-3.12.1.tar.bz2 && \
cd geos-3.12.1 && \
mkdir _build && cd _build && \
cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/usr/local .. && \
make -j12 && ctest && make install

FLANN 1.9.2

cd ~ && \
curl -L https://github.com/flann-lib/flann/archive/refs/tags/1.9.2.tar.gz > flann.tar.gz && \
tar xf flann.tar.gz && \
cd flann-1.9.2 && \
mkdir build && cd build && \
cmake -DCMAKE_INSTALL_PREFIX=/usr/local -DBUILD_C_BINDINGS=OFF -DBUILD_PYTHON_BINDINGS=OFF \
      -DBUILD_MATLAB_BINDINGS=OFF -DBUILD_EXAMPLES=OFF -DBUILD_TESTS=OFF -DBUILD_DOC=OFF .. && \
make -j12 && make install

nanoflann v1.6.2

cd ~ && \
curl -L https://github.com/jlblancoc/nanoflann/archive/refs/tags/v1.6.2.tar.gz > nanoflann.tar.gz && \
tar xf nanoflann.tar.gz && \
cd nanoflann-1.6.2 && \
mkdir build && cd build && \
cmake -DCMAKE_INSTALL_PREFIX=/usr/local .. && \
make -j12 && make install

Armadillo 12.8.4

sudo apt install -y cmake libopenblas-dev liblapack-dev libarpack2-dev libsuperlu-dev && \
curl -L https://sourceforge.net/projects/arma/files/armadillo-12.8.4.tar.xz > ~/armadillo-12.8.4.tar.xz && \
cd ~ && tar -xf ~/armadillo-12.8.4.tar.xz && cd ~/armadillo-12.8.4 && \
cmake . && \
sudo make install && \
cd ~ && rm -rf ~/armadillo-12.8.4.tar.xz && rm -rf ~/armadillo-12.8.4