Voronoi.Based.Hybrid.A.mp4
The following project is an extension of my bachelor's thesis conducted under the guidance of Prof. Arpita Sinha at Indian Institute of Technology Bombay (IITB) and submitted at Charotar University of Science and Technology. The goal of the project was to implement a suitable path planning strategy for a tractor-trailer system. In our implementation, we extend the Hybrid A* algorithm, first by introducing novelties to the heuristic search to suit the tractor-trailer model and then present a modified analytical expansion step based on voronoi graph to guide the search towards the goal faster.
The code in this repository is specifically tested on Ubuntu 18.04 and ROS Melodic. The implementation is based on a Clearpath Robotics Ridgeback robot towing a custom designed trailer, however it can be tuned to other tractor-trailer systems easily.
- Ubuntu 18.04
- ROS Melodic
- Open Motion Planning Library (OMPL)
- jsk_reconition_msgs
- tuw_multi_robot Follow INSTALL.md given in the repository. Refer to this issue if compilation error exists. Install it in the same catkin workspace as Voronoi based Hybrid A*.
sudo apt install libompl-dev
sudo apt-get install ros-melodic-jsk-recognition-msgs
mkdir -p catkin_ws/src
cd catkin_ws/src
git clone https://github.com/tanujthakkar/Voronoi-Based-Hybrid-Astar.git
cd ..
catkin_make
source devel/setup.bash
roslaunch hybrid_astar hybrid_astar.launch
Wait for the voronoi graph to be computed and visualized in RViz. After that, 2D Pose Estimate and 2D Nav Goal tools can be used to specify start and goal positions respectively for the planner.
We are currently working on a paper which will soon be added here to cite.