This project focuses on path planning of non-holonomic robots in dynamic environments. For the environment amazon warehouse environment is chosen in gazebo and created a 2d map of it. For global planners AIT* and BIT* are chosen and compared with RRT. For local planner MPC is used and compared with APF. Gazebo simulation is done with ROS2 Humble.
To run the global planners in 2d, download the files from global planners folder and run the following commands:
- python3 batch_informed_trees.py for BIT*
- python3 adaptively_informed_trees_star.py for AIT*
- python3 rrt.py for RRT
To run the MPC in 2d, download the file from folder local planner and run python3 MPC.py
To run MPC in gazebo, download the src folder. Go to the directory where src is downloaded and run colcon build in terminal. After the package is build, run the following commands from the root folder(where src is placed) for completing the setup. These commands should be run everytime before running the simulation.
- . install/local_setup.bash
- export TURTLEBOT3_MODEL=burger
- export GAZEBO_MODEL_PATH=$GAZEBO_MODEL_PATH:`ros2 pkg
prefix turtlebot3_gazebo
`/share/turtlebot3_gazebo/models/
After completing the setup run the following command for spawning the robot and world in gazebo, ros2 launch gazbeosim world_with_robot_new.launch.py. Then run the following command for runing MPC in gazebo ros2 run planner MPC.