Hands-on with TurtleBot3 — Humble → Jazzy conversion

Replacements for each slide in Hands_on_TurtleBot3.pdf, targeting ROS 2 Jazzy Jalisco on Ubuntu 24.04 with Gazebo Harmonic.

---

What actually changed

Area	Humble	Jazzy
Distro path	/opt/ros/humble	/opt/ros/jazzy
Git branch	-b humble	-b jazzy
Simulator	Gazebo Classic 11 (gazebo_ros)	Gazebo Harmonic (ros_gz)
/cmd_vel type	geometry_msgs/Twist	geometry_msgs/TwistStamped
RViz config	hand-built path into turtlebot3_gazebo	turtlebot3_bringup rviz2.launch.py
TF tool	view_frames.py	view_frames
ros2 pkg create	license optional	warns unless --license given

Launch file names are unchanged (turtlebot3_world.launch.py, navigation2.launch.py), so most of your ros2 launch lines survive as-is. The three that need real edits are Step 0, the controller script, and the RViz2 slide.

---

Slide 4 — Step 0: Prepare Workspace & Dependencies

Assume ROS 2 Jazzy + Gazebo Harmonic are already installed (Ubuntu 24.04).

mkdir -p ~/ros2_ws/src
cd ~/ros2_ws/src

Source Jazzy:

source /opt/ros/jazzy/setup.bash

Install the simulation and navigation dependencies:

sudo apt update
sudo apt install -y \
  ros-jazzy-ros-gz \
  ros-jazzy-navigation2 ros-jazzy-nav2-bringup \
  ros-jazzy-cartographer-ros \
  python3-colcon-common-extensions python3-rosdep

ros-jazzy-ros-gz pulls in Gazebo Harmonic. Gazebo Classic is not available on Ubuntu 24.04 — it reached end of life in January 2025, which is why the simulator stack changed.

---

Slide 5 — Step 1: Clone TurtleBot3 Repositories

Clone Core Packages

cd ~/ros2_ws/src
# Core TurtleBot3 packages (Jazzy branch)
git clone -b jazzy \
  https://github.com/ROBOTIS-GIT/turtlebot3.git
git clone -b jazzy \
  https://github.com/ROBOTIS-GIT/turtlebot3_msgs.git

Clone Simulation Package

# Gazebo Harmonic simulation for TurtleBot3
git clone -b jazzy \
  https://github.com/ROBOTIS-GIT/turtlebot3_simulations.git

Optional — real hardware only

git clone -b jazzy \
  https://github.com/ROBOTIS-GIT/DynamixelSDK.git

Result: turtlebot3, turtlebot3_msgs, and turtlebot3_simulations are now in ~/ros2_ws/src.

Shortcut for a lab full of machines: Jazzy binaries exist, so sudo apt install ros-jazzy-turtlebot3 ros-jazzy-turtlebot3-msgs ros-jazzy-turtlebot3-gazebo skips the clone-and-build entirely. Worth checking with apt search ros-jazzy-turtlebot3 before the session — it saves students a 10-minute build.

---

Slide 6 — Step 2: Build & Source the Workspace

Resolve dependencies first

cd ~/ros2_ws
rosdep install --from-paths src --ignore-src -r -y

Build with colcon

cd ~/ros2_ws
colcon build --symlink-install

Source the Overlay

source ~/ros2_ws/install/setup.bash

• Repeat the source command in every new terminal (or add it to ~/.bashrc).
• Set the model once in ~/.bashrc so students stop forgetting it:

  echo 'export TURTLEBOT3_MODEL=burger' >> ~/.bashrc

---

Slides 8–13 — ROS 2 Launch Files

No changes needed. LaunchDescription, launch_ros.actions.Node, generate_launch_description(), and the setup.py data_files glob all behave identically in Jazzy.

One small addition for the setup.py slide (slide 13) — Jazzy is stricter about package metadata, so make sure setup.py also carries a license field:

# Add to imports:
from glob import glob
import os

# Add to data_files in setup():
(os.path.join('share', package_name, 'launch'), glob('launch/*.py')),

# And in setup():
license='Apache-2.0',

---

Slide 15 — Step 3: Launch Simulation

Start Gazebo Harmonic Simulation Environment

# Choose the robot model
export TURTLEBOT3_MODEL=burger

# Launch TurtleBot3 World in Gazebo Harmonic
ros2 launch turtlebot3_gazebo turtlebot3_world.launch.py

Command is unchanged, but the window that opens is the new Gazebo (Harmonic), not Gazebo Classic. First launch takes longer while resources are fetched.

Other worlds:

ros2 launch turtlebot3_gazebo empty_world.launch.py
ros2 launch turtlebot3_gazebo turtlebot3_house.launch.py

---

Slide 16 — Step 4: Explore Available ROS 2 Topics

ros2 topic list

Expected topics are the same, but the type of /cmd_vel has changed:

• /cmd_vel — Velocity commands (input) → now geometry_msgs/msg/TwistStamped
• /odom — Odometry data (output)
• /scan — Laser scan data
• /imu — Inertial measurement unit data
• /joint_states — Robot joint states
• /tf and /tf_static — Transform data

Good live demo to add here:

ros2 topic type /cmd_vel
ros2 interface show geometry_msgs/msg/TwistStamped

---

Slide 17 — Step 5: Teleoperate the Robot

ros2 run turtlebot3_teleop teleop_keyboard

Command and key bindings (W X A D S) are unchanged. Internally the node now publishes TwistStamped.

---

Slide 18 — Step 6: Monitor Topic Activity

ros2 topic echo /cmd_vel
ros2 topic echo /odom

rqt_graph
rqt

Unchanged. Note that ros2 topic echo /cmd_vel output now shows a header block above twist — a natural place to point out the message change.

---

Slide 19 — Step 7: Launch RViz2

The old ros2 pkg prefix turtlebot3_gazebo/.rviz path is fragile on Jazzy. Use the bringup launch file instead:

# Launch RViz2 with the TurtleBot3 configuration
ros2 launch turtlebot3_bringup rviz2.launch.py

# Or launch bare and add displays manually
ros2 run rviz2 rviz2

Key displays to add (unchanged): TF, LaserScan (/scan), RobotModel, Odometry, PointCloud2.

---

Slide 20 — RViz2 Configuration Tips

Unchanged, with one rename: the goal tool in the Jazzy toolbar is labelled "2D Goal Pose" (the Nav2 plugin also exposes it as "Nav2 Goal"), not "2D Nav Goal".

• Global Options → Fixed Frame: odom
• LaserScan → Topic /scan, Size 0.1 m
• Interactive Markers → 2D Pose Estimate, 2D Goal Pose

---

Slide 21 — Step 8: Additional Exploration Commands

# View all running nodes
ros2 node list
# View node information
ros2 node info /teleop_keyboard
# View topic message structure
ros2 interface show sensor_msgs/msg/LaserScan
# Monitor specific topic rate
ros2 topic hz /scan
# View TF tree  (note: no .py suffix in Jazzy)
ros2 run tf2_tools view_frames

---

Slide 23 — Step 1: Create a ROS 2 Package

cd ~/ros2_ws/src
ros2 pkg create turtlebot3_controller \
  --build-type ament_python \
  --license Apache-2.0 \
  --dependencies rclpy geometry_msgs

--license is new in practice: Jazzy's ros2 pkg create emits a warning without it, and colcon test complains later.

---

Slide 24 — Step 2: Controller Script (rewritten for Jazzy)

Create turtlebot3_controller/circle_controller.py.

Three things changed from the Humble version:

1. Twist → TwistStamped (velocity now lives under msg.twist, plus a header).
2. Publishing once inside __init__ never worked reliably — the publisher has not finished discovery yet, so the single message is dropped. Use a timer.
3. linear.x = 1.0 exceeds the burger's maximum of 0.22 m/s; it gets clamped and the circle is not the radius you advertise.

import rclpy
from rclpy.node import Node
from geometry_msgs.msg import TwistStamped


class Circle(Node):
    def __init__(self):
        super().__init__('turtlebot3_controller')
        self.get_logger().info('Node started with turtlebot3_circle')

        self.vel_pub = self.create_publisher(TwistStamped, '/cmd_vel', 10)

        self.linear_x = 0.15    # m/s   (burger max: 0.22 m/s)
        self.angular_z = 0.5    # rad/s (burger max: 2.84 rad/s)

        radius = self.linear_x / self.angular_z
        self.get_logger().info(f'Radius = {radius:.2f} m')

        self.timer = self.create_timer(0.1, self.publish_velocity)  # 10 Hz

---

Slide 25 — Step 2 continued

    def publish_velocity(self):
        msg = TwistStamped()
        msg.header.stamp = self.get_clock().now().to_msg()
        msg.header.frame_id = 'base_link'
        msg.twist.linear.x = self.linear_x
        msg.twist.angular.z = self.angular_z
        self.vel_pub.publish(msg)

    def stop(self):
        msg = TwistStamped()
        msg.header.stamp = self.get_clock().now().to_msg()
        msg.header.frame_id = 'base_link'
        self.vel_pub.publish(msg)


def main(args=None):
    rclpy.init(args=args)
    node = Circle()
    try:
        rclpy.spin(node)
    except KeyboardInterrupt:
        pass
    finally:
        node.stop()
        node.destroy_node()
        if rclpy.ok():
            rclpy.shutdown()


if __name__ == '__main__':
    main()

• Publisher: publishes TwistStamped velocity commands to /cmd_vel.
• Velocity setup: twist.linear.x controls forward speed, twist.angular.z the rotation rate.
• Radius: r = v / ω, so 0.15 / 0.5 = 0.3 m.
• Stopping on Ctrl+C matters — otherwise the robot keeps circling after the node dies.

---

New slide worth adding — setup.py entry point

The original deck never shows this, but ros2 run turtlebot3_controller circle_controller fails without it:

entry_points={
    'console_scripts': [
        'circle_controller = turtlebot3_controller.circle_controller:main',
    ],
},

---

Slide 26 — Step 3: Build the Package

cd ~/ros2_ws
colcon build --symlink-install --packages-select turtlebot3_controller

source ~/ros2_ws/install/setup.bash

Unchanged.

---

Slide 27 — Step 4: Launch Simulation and Run Controller

export TURTLEBOT3_MODEL=burger
ros2 launch turtlebot3_gazebo turtlebot3_world.launch.py

ros2 run turtlebot3_controller circle_controller

Result: The TurtleBot3 follows a circular path of radius ≈ 0.3 m in Gazebo Harmonic.

---

Slide 31 — Nav2 Step 1: Navigation Setup

Simulation Environment

# Launch Gazebo Harmonic simulation
export TURTLEBOT3_MODEL=burger
ros2 launch turtlebot3_gazebo turtlebot3_world.launch.py

Real Robot Environment

# Launch navigation with pre-built map
ros2 launch nav2_bringup bringup_launch.py \
  use_sim_time:=false \
  map:=$HOME/tb3_world_map.yaml

• Simulation: use_sim_time:=True for the Gazebo clock
• Real robot: use_sim_time:=false for the system clock
• Map path: ensure the path to your saved map file is correct

You need a map before this step. Build one with:

ros2 launch turtlebot3_cartographer cartographer.launch.py use_sim_time:=True
ros2 run nav2_map_server map_saver_cli -f ~/map

---

Slide 32 — Nav2 Step 2: Launch Navigation Stack

export TURTLEBOT3_MODEL=burger
ros2 launch turtlebot3_navigation2 navigation2.launch.py \
  use_sim_time:=True \
  map:=$HOME/map.yaml

Components launched (unchanged): AMCL localization, global + local planners, DWB controller, behavior trees.

Add this note to the slide: Nav2 on Jazzy publishes TwistStamped on /cmd_vel by default. If you are mixing in a node that expects plain Twist, set enable_stamped_cmd_vel: false in the parameter files of both turtlebot3_bringup and turtlebot3_navigation2. Keeping the whole stack on one convention avoids a silent "robot never moves" failure.

---

Slides 33–34 — Set Initial Pose and Navigation Goal

Substantively unchanged. Only the toolbar label for the goal tool differs:

• 2D Pose Estimate — same name, same behaviour
• 2D Goal Pose — this was "2D Nav Goal" in the Humble-era RViz2

Troubleshooting notes on both slides still apply.

---

Slide 35 — Monitor Robot Motion

Unchanged. rqt tools (Topic Monitor, Node Graph, Bag Recording, Plotting, Console, Robot Steering) and the RViz2 displays (Map, Path, LaserScan, TF, Costmaps) all work the same.

One caveat for Robot Steering in rqt: it publishes plain Twist, so on Jazzy it will not drive the robot unless you have switched the stack to enable_stamped_cmd_vel: false. Either mention it or drop it from the list.