This meta-package is for 2D-map exploration for wheeled robots based on Nav2 stack. The exploration algorithm is based on frontier searching. For motion planning TEB local planner is used.
You need to pull this docker image. After that, use run.bash script to run a container and run a new bash session
by using into.bash
docker pull andrewfram/map_exploration:latest
cd map_exploration
bash docker/run.bash
bash docker/into.bash
Inside the docker container build the packages and source the environment:
cd workspace
colcon build
source ~/.bashrc
In two bash sessions run the following commands to launch the gazebo environment and exploration algorithm. The TB3 should start exploration:
ros2 launch survey maze_slam.launch.py
ros2 run frontier_discoverer discoverer
In the survey package there are configuration files that set parameters for planning and mapping algorithms. The
most important parameters to adjust are provided under controller_server tag in nav2_params_tb3.yaml. If the
robot gets stuck you might need to change the mapping resolution in slam_toolbox_tb3.yaml.