Module 03 · ROS 2

Almost every modern robot is a set of programs working together: one reads sensors, one plans, one controls motors, one watches for danger. ROS 2 (the Robot Operating System, version 2) is the open-source toolkit that lets those programs find each other, share data and be started, inspected and replayed in a consistent way. Learning it is the biggest single step from "I can code" to "I can build a robot system", and it is used in research labs and companies worldwide.

This module teaches ROS 2 Jazzy Jalisco on Ubuntu 24.04, in Python. You start with how ROS 2 is put together, then write real packages: publishers, subscribers, services, actions, parameters and launch files. You describe a robot with URDF, keep track of its coordinate frames with TF2, and learn the debugging and recording tools that professionals use daily. Everything runs on your own machine and needs no hardware.

By the end of this module you can…

• Explain the ROS 2 graph, DDS discovery, QoS and executors, and pick sensible QoS settings.
• Write, build and run Python packages with publishers, subscribers, services and clients.
• Use actions for long-running tasks, parameters for configuration, and launch files for whole systems.
• Describe a two-wheeled robot in xacro, publish its transform tree and view it in RViz.
• Debug a live system with the ros2 CLI and rqt, and record and replay data with rosbag2.

Lessons

Lesson	What you build	Time
ROS 2 architecture and DDS	A latched publisher, a late subscriber and a two-speed executor demo	40 min
Nodes, topics and services	The kinemata_basics package: talker, listener, add service, turtle driver	45 min
Actions, parameters and launch files	A Fibonacci action, a configurable speed limiter and a launch file	45 min
TF2 and URDF	A TF broadcaster and listener, and a two-wheeled robot in RViz	45 min
Packages, debugging and bags	A clean workspace build, a recorded bag and a bag reader	40 min

Module project: the Kinemata base

Pull the pieces together into one launch file called kinemata_base.launch.py. It should start robot_state_publisher with your xacro model, a node that turns cmd_vel into a simulated odometry broadcast on odom → base_link, the speed limiter with a launch-argument limit, and RViz. Drive it with ros2 topic pub or your circle driver, and record a bag of /tf, /tf_static and /cmd_vel for 30 seconds. Then replay the bag with nothing else running and confirm the robot still moves in RViz. When that works you have a complete, inspectable, reproducible ROS 2 system, ready to be placed in a simulated world in the next module.

Start with ROS 2 architecture and DDS.