How joshnewans/articubot_one works
Project type and purpose
• This is a ROS2 robot package template that unifies:
• Robot description (URDF/XACRO)
• Sensor and actuator configurations (Gazebo, ROS2 Control)
• Parameter files for mapping, navigation, controllers, trackers
• Launch files for both simulation and real hardware
• Its goal is to let users clone and immediately run a robot in Gazebo, then deploy on real hardware, swapping configs.
File structure
• Top-level:
• CMakeLists.txt & package.xml – build / package metadata
• config/ – YAML parameter sets for robot controllers, navigation (Nav2), trackers, RViz, joystick
• description/ – XACRO macros and URDF fragments for robot body, sensors (camera, lidar), control plugins
• launch/ – Python-based ROS2 launch scripts (simulation vs. hardware, mapping, navigation, ball tracker, joystick)
• worlds/ – Gazebo world definitions (empty, obstacles)
• Patterns:
• Modular description: each sensor in its own XACRO
• Separate launch files for each major use case (sim vs real, perception vs navigation)
• Parameterization by YAML for easy swapping
Extract insights from the README
• It’s intended as a template—rename instances of “my_bot”
• Directories can be pruned once you adapt your own robot
Main components & their relationships
a. Components:
1. Robot Description Layer
– XACRO files → URDF model
2. Simulation / Hardware Layer
– Gazebo + world file + robot Gazebo plugins
– Real hardware interface via ros2_control
3. ROS2 Control Manager
– Reads config/my_controllers.yaml
– Spawns controllers for wheels, joints
4. Sensor Nodes & Drivers
– Depth camera node (camera.xacro + camera.launch.py)
– LIDAR node (rplidar.launch.py)
5. Perception & Tracking
– ball_tracker.launch.py (tracks ball pose)
6. Teleoperation
– joystick.launch.py → publishes Twist commands
7. Mapping & Localization
– mapper_params_online_async.yaml + online_async_launch.py
– localization_launch.py
8. Navigation (Nav2)
– navigation_launch.py + nav2_params.yaml
9. Visualization
– RViz config files in config/…rviz
b. Data flows & interactions
– The Gazebo plugin publishes sensor topics (/camera/image, /scan) into the ROS graph
– Sensor nodes publish into TF tree and topic streams
– Perception node (/ball_tracker) subscribes to image topics, publishes ball pose
– Mapping node subscribes to /scan, /odom → publishes map, pose
– Localization node fuses odom + map → /amcl_pose
– Navigation node subscribes to /amcl_pose + /goal → publishes /cmd_vel to ROS2 Control
– Teleop node can also publish /cmd_vel, gated by twist_mux
– Controllers subscribe to /cmd_vel, send actuator commands via ros2_control
c. Architectural patterns & principles
– Modular XACRO composition for reusability
– Clear environment separation: simulation vs hardware via launch files
– Parameter-driven design: all tuning resides in config/*.yaml
– ROS2 node-based microservices communicating over well-defined topics
d. Key technologies
– ROS2 Foxy (or later)
– Gazebo (physics, sensor plugins)
– ros2_control + controller_manager
– Nav2 (navigation stack)
– AMCL (localization), SLAM (online async)
– RViz2 for visualization
– Python-based launch system