RRT* with Reeds-Shepp Curve

Introduction

RRT* is an asymptotically optimal path planning algorithm. The planned path can meet the kinetics requirements of the car when using RS curve to connect the sampling points. But there is currently no suitable open source project developed based on C++ and ROS, this project is open source to help those who need it.

Project file description:

• app/search.cpp is the main file of node.
• Generate maps using OpenCV, and map data is saved by csv file in benchmarks/.
• The configuration files for each scene are stored in config_yaml/.
• map/ and map_yaml/ store visualization files for rviz.
• velocity_set.cpp performs post-processing on the generation path, including sampling, visualization, saving, etc.

Demo:

1. case1:

2. case2:

Dependencies:

• ROS. Melodic or Noetic can be satisfied.
• Eigen3
• glog
• yaml-cpp
• matplotlib-cpp(included in the ./include/, no need to install separately)
• openCV

Build:

# in your ros workspace, for example: your_ws/src
git clone git@github.com:reflector-li/rrt_star_rs.git
cd ..
catkin_make

Usage

# in your ros workspace, for example: your_ws/
source devel/setup.bash
roslaunch rrt_star_rs rrt_search.launch

Reference

• Hybrid_A_Star

  The code framework and obstacle detection strategy of this project refer to it.

• PythonRobotics