This repository contains a world and launch files for testing the the Multi-Robot Graph SLAM Framework in a simulated environment using Gazebo (tested on ROS2 humble and Gazebo Fortress).
This package depends on the following packages:
sudo apt install ros-humble-ros-gz
The demo explained in this section can be checked out in this video on youtube:
Launch the world in gazebo with the following command. Note that the first call to gazebo might take a while, since it has to download the models of the fortress.
ros2 launch mrg_slam_sim marsyard2020.launch.py
Spawn two robots equipped with LIDAR sensors and bridge the sensor data to ROS2 topics. Also start the teleop_joystick node to control the first robot atlas with a joystick (xbox). This terminal has to be kept active.
ros2 launch mrg_slam_sim dual_robot_sim.launch.py
You should be able to control the first robot atlas with the joystick.
If you don't have a joystick, you can disable the teleop_joystick node in the dual robot config by setting enable_teleop_joy to false. Then you can control the robot with the keyboard using the following command:
ros2 run teleop_twist_keyboard teleop_twist_keyboard --ros-args -r __node:=teleop_twist_keyboard_node_atlas -r /cmd_vel:=/atlas/cmd_vel
Second robot bestla can be controlled with the keyboard in a new terminal using the following command:
ros2 run teleop_twist_keyboard teleop_twist_keyboard --ros-args -r __node:=teleop_twist_keyboard_node_bestla -r /cmd_vel:=/bestla/cmd_vel
Now to test the Multi-Robot Graph SLAM Framework with the mrg_slam package. Run two instances of the slam algorithm as follows:
ros2 launch mrg_slam mrg_slam.launch.py model_namespace:=atlas x:=-15.0 y:=13.5 z:=1.2 # terminal 1
ros2 launch mrg_slam mrg_slam.launch.py model_namespace:=bestla x:=-15.0 y:=-13.0 z:=1.2 # terminal 2
Or use two instances of Multi-Robot Graph SLAM inside two docker containers. Make sure that the docker containers can communicate with the host machine. For example, environment variables like ROS_LOCALHOST_ONLY or ROS_DOMAIN_ID should not be set or should be correctly set. Then run the following commands:
docker run -it --rm --network=host --ipc=host --pid=host -e MODEL_NAMESPACE=atlas -e USE_SIM_TIME=true -e X=-15.0 -e Y=13.5 -e Z=1.2 --name atlas_slam mrg_slam # terminal 1
docker run -it --rm --network=host --ipc=host --pid=host -e MODEL_NAMESPACE=bestla -e USE_SIM_TIME=true -e X=-15.0 -e Y=-13.0 -e Z=1.2 --name bestla_slam mrg_slam # terminal 2
Move the robots and visualize the results in rviz2:
rviz2 -d /path/to/mrg_slam/rviz/mrg_slam.rviz --ros-args -p use_sime_time:=true
Many packages use hard-coded frames like /base_link or /odom. If you want to test the Multi-Robot Graph SLAM Framework with a single robot without a namespace, you can use the single_robot_sim.launch.py launch file. After launching the gazebo world in step 1 above, you can spawn a single robot.
ros2 launch mrg_slam_sim single_robot_sim.launch.py
If you want to test the Multi-Robot Graph SLAM Framework with a single robot with a namespace, you need to set the robot_name parameter in the single_robot_marsyard2020_sim.yaml file. After launching the gazebo world in step 1 above, you can spawn a single robot with a namespace.
ros2 launch mrg_slam_sim single_robot_sim.launch.py
At the moment there are two worlds available: marsyard2020 and rubicon. You can launch the rubicon world with the following command:
ros2 launch mrg_slam_sim rubicon.launch.py
To spawn two robots at the correct positions in the rubicon world, you can use the following command:
ros2 launch mrg_slam_sim dual_robot_sim.launch.py world:=rubicon # dual robot in rubicon world
ros2 launch mrg_slam_sim single_robot_sim.launch.py world:=rubicon # single robot in rubicon world
Note that when you start an instance of the Multi-Robot Graph SLAM Framework, you have to set the correct initial position of the robot in the rubicon world. The initial poses can be found in the config folder.
ros2 launch mrg_slam mrg_slam.launch.py model_namespace:=atlas x:=-7.0 y:=0.0 z:=3.96 # terminal 1
ros2 launch mrg_slam mrg_slam.launch.py model_namespace:=bestla x:=0.0 y:=-15.0 z:=3.8 # terminal 2