Mahmoud Abas, Full-Stack Software Engineer & AI

ROS 2 Practical - Lab: Simulating a Robot Using Gazebo Jetty and ROS 2 Jazzy

In the print dialog, choose "Save as PDF" as the destination.

Lab: Simulating a Robot Using Gazebo Jetty and ROS 2 Jazzy

Building and bridging a warehouse robot in simulation

Learning Objectives

Understand Gazebo Classic vs Gazebo Jetty (gz-sim) differences
Create robot models using SDF format
Use gz-sim plugins for control and sensors
Integrate ROS 2 Jazzy with Gazebo Jetty via ros_gz
Bridge Gazebo topics to ROS 2 topics
Build ROS 2 launch files for simulation

Critical

Gazebo Jetty vs Classic

Feature	Classic	Jetty (New)
Command	`gazebo`	`gz sim`
Package	`gazebo_ros_pkgs`	`ros_gz`
Plugins	`libgazebo_ros_*.so`	`gz-sim-*-system`
Model path	`GAZEBO_MODEL_PATH`	`GZ_SIM_RESOURCE_PATH`
DiffDrive	`libgazebo_ros_diff_drive.so`	`gz-sim-diff-drive-system`

Install and Verify

bash

sudo apt-get install gz-jetty
sudo apt install ros-jazzy-ros-gz   # bridge + sim + image packages

# Verify
ros2 doctor
gz sim --version

Launching and Exploring Gazebo

bash

gz sim shapes.sdf      # GUI
gz sim -s               # server only (headless)
gz sim -g               # GUI only
gz sim -v 4              # verbose

Key GUI plugins: Entity Tree, 3D View, World Control, Transform Control
Every world needs a Physics and SceneBroadcaster system plugin

Building the Robot Model

Model lives in ~/.gz/models/mobile_warehouse_robot/, not ~/.gazebo/models
model.config declares name, version, and points to model.sdf
model.sdf defines links (base, left/right wheel, caster, lidar) with inertial, visual, and collision blocks

xml

<link name="base_link">
  <inertial><mass>15.0</mass>...</inertial>
  <visual name="base_visual">...</visual>
  <collision name="base_collision">...</collision>
</link>

Wheels, Caster, and Lidar Sensor

base_link

main body

left_wheel

revolute joint

right_wheel

revolute joint

caster

ball joint

lidar_link

fixed joint

Left/right wheels: revolute joints, cylinder geometry, friction surface
Caster: ball joint, sphere geometry, low friction (free rolling)
GPU lidar sensor attached via a fixed joint on lidar_link

xml

<sensor name="gpu_lidar" type="gpu_lidar">
  <topic>lidar</topic>
  <update_rate>10</update_rate>
  <lidar>
    <scan><horizontal><samples>640</samples>
      <min_angle>-1.57</min_angle><max_angle>1.57</max_angle>
    </horizontal></scan>
    <range><min>0.08</min><max>10.0</max></range>
  </lidar>
</sensor>

Differential Drive Plugin

Joints connect wheels to base_link; the DiffDrive plugin drives them together

xml

<plugin filename="gz-sim-diff-drive-system"
        name="gz::sim::systems::DiffDrive">
  <left_joint>left_wheel_joint</left_joint>
  <right_joint>right_wheel_joint</right_joint>
  <wheel_separation>0.5</wheel_separation>
  <wheel_radius>0.08</wheel_radius>
  <topic>/model/warehouse_robot/cmd_vel</topic>
  <odom_topic>/model/warehouse_robot/odometry</odom_topic>
</plugin>

Building the Warehouse World

Separate small_warehouse model plus a top-level warehouse.sdf world file
World includes: physics, required system plugins (Physics, UserCommands, SceneBroadcaster, Sensors), GUI panels, lighting, ground plane
Models referenced with model:// URIs and placed via <include><pose>...

xml

<include>
  <name>warehouse_robot</name>
  <pose>-4 -4 0.1 0 0 0</pose>
  <uri>model://mobile_warehouse_robot</uri>
</include>

Bridging Gazebo to ROS 2

ros_gz_bridge: Translates between Gazebo Transport (gz.msgs.*) and ROS 2 message types, defined declaratively in a bridge.yaml.

ROS 2

geometry_msgs, nav_msgs...

/cmd_vel

ros_gz_bridge

bridge.yaml

Gazebo Transport

gz.msgs.*

yaml

- ros_topic_name: "/cmd_vel"
  gz_topic_name: "/model/warehouse_robot/cmd_vel"
  ros_type_name: "geometry_msgs/msg/Twist"
  gz_type_name: "gz.msgs.Twist"
  direction: ROS_TO_GZ

- ros_topic_name: "/odom"
  gz_type_name: "gz.msgs.Odometry"
  direction: GZ_TO_ROS

Bridges /cmd_vel in, and /odom, /scan, /tf out

ament_python gotcha

Package Layout

text

warehouse_simulation/
├── package.xml
├── setup.py
├── launch/warehouse_simulation.launch.py
└── config/bridge.yaml

setup.py must glob() these into data_files so colcon installs them

Launch File

One launch file starts Gazebo and the bridge together

python

gz_sim = IncludeLaunchDescription(
    PythonLaunchDescriptionSource(
        os.path.join(pkg_ros_gz_sim, 'launch', 'gz_sim.launch.py')),
    launch_arguments={'gz_args': f'-r {world_file}'}.items())

bridge = Node(
    package='ros_gz_bridge',
    executable='parameter_bridge',
    arguments=['--ros-args', '-p', f'config_file:={bridge_config}'])

Build, Launch, and Verify

Build

colcon build

Source

install/setup.bash

Launch

ros2 launch

Verify

ros2 topic list

bash

colcon build --packages-select warehouse_simulation
source install/setup.bash

ros2 launch warehouse_simulation warehouse_simulation.launch.py

In a new terminal, confirm the bridge is live:

bash

ros2 topic list
# /cmd_vel  /odom  /scan  /tf ...

Key Takeaways

Gazebo Jetty (gz sim + ros_gz + gz-sim-*-system plugins) replaces Gazebo Classic entirely, do not mix tutorials
Robot models are SDF: links (inertial/visual/collision) plus joints plus plugins
Absolute topic paths in the SDF are mandatory for the bridge to match names correctly
ros_gz_bridge plus a YAML config is the standard way to connect Gazebo topics to ROS 2 topics
ament_python packages need launch/ and config/ at the package root, wired into setup.py's data_files
One launch file typically starts both Gazebo and the bridge together

Open this lesson

Mahmoud AbasLab: Simulating a Robot Using Gazebo Jetty and ROS 2 Jazzy

Feature

Classic

Jetty (New)

Command

gazebo

gz sim

Package

gazebo_ros_pkgs

ros_gz

Plugins

libgazebo_ros_*.so

gz-sim-*-system

Model path

GAZEBO_MODEL_PATH

GZ_SIM_RESOURCE_PATH

DiffDrive

libgazebo_ros_diff_drive.so

gz-sim-diff-drive-system

<sensor name="gpu_lidar" type="gpu_lidar"> <topic>lidar</topic> <update_rate>10</update_rate> <lidar> <scan><horizontal><samples>640</samples> <min_angle>-1.57</min_angle><max_angle>1.57</max_angle> </horizontal></scan> <range><min>0.08</min><max>10.0</max></range> </lidar> </sensor>

<plugin filename="gz-sim-diff-drive-system" name="gz::sim::systems::DiffDrive"> <left_joint>left_wheel_joint</left_joint> <right_joint>right_wheel_joint</right_joint> <wheel_separation>0.5</wheel_separation> <wheel_radius>0.08</wheel_radius> <topic>/model/warehouse_robot/cmd_vel</topic> <odom_topic>/model/warehouse_robot/odometry</odom_topic> </plugin>

- ros_topic_name: "/cmd_vel" gz_topic_name: "/model/warehouse_robot/cmd_vel" ros_type_name: "geometry_msgs/msg/Twist" gz_type_name: "gz.msgs.Twist" direction: ROS_TO_GZ - ros_topic_name: "/odom" gz_type_name: "gz.msgs.Odometry" direction: GZ_TO_ROS

gz_sim = IncludeLaunchDescription( PythonLaunchDescriptionSource( os.path.join(pkg_ros_gz_sim, 'launch', 'gz_sim.launch.py')), launch_arguments={'gz_args': f'-r {world_file}'}.items()) bridge = Node( package='ros_gz_bridge', executable='parameter_bridge', arguments=['--ros-args', '-p', f'config_file:={bridge_config}'])