multi_turtlebot_sim

Introduction

The credit goes to Franco Cipollone who created this package for Turtlebot3 Waffle model. I just edited and added files for it to make it work for turtlebot3 Burger model. This package is meant to be an example on how robots needs to be configured in order to support simulating multi robots in the same environment.

Platform

ROS 2 Version: Foxy

Notes

TF

• One /tf topic is expected to be published for any number of robots
• Each frame should have a prefix to know which robot they belong to

ROS

• Topics should be namespaced.
  • <robot_prefix>/robot_description
  • <robot_prefix>/joint_states
  • <robot_prefix>/cmd_vel
  • <robot_prefix>/odom
  • <robot_prefix>/scan
• Nodes should be namespaced.
  • robot_state_publisher: Reads robot description and:
    • Publishes <robot_prefix>/robot_description topic.
    • Publishes transforms to /tf. Transforms should be prefixed for the robot.
    • Subscribes to joint_states. Topic should be prefixed for the robot.
  • Gazebo related nodes
    • gazebo_ros_diff_drive plugin: Should be namespaced.
    • gazebo_ros_joint_state_publisher plugin: Should be namespaced.

Try it!

• Launch the standalone simulation, no robot is spawned

ros2 launch multi_burger_sim standalone_world.launch.py

Note: By default the empty world is used, you can also use world_name:=turtlebot3_world.world.

• Spawn a robot using a robot_prefix in a particular x_pose and y_pose.

ros2 launch multi_burger_sim spawn_turtlebot3.launch.py robot_prefix:=burger1 x_pose:=0.5 y_pose:=0.5

The robot_prefix will affect both the ros namespace and the tf prefix for that robot

• Spawn another robot in a different position

ros2 launch multi_burger_sim spawn_turtlebot3.launch.py robot_prefix:=burger2 x_pose:=-0.5 y_pose:=-0.5

Check how the tf tree is conformed:

ros2 run tf2_tools view_frames.py

• The tf tree of each robot aren't connected therefore both robots in rviz can't be seen. As the simulation is starting the odom frame for each robot in the {0,0} world position we can add a static transform between both odometry frames to achieve the full tf_tree all connected.

ros2 run tf2_ros static_transform_publisher 0 0 0 0 0 0 burger1/odom burger2/odom