Gazebo documentation

Learn how to use Gazebo Harmonic for realistic robot simulation with physics, sensors, and 3D visualization.

Official documentation

Complete simulator reference including API documentation, tutorials, and migration guides

About Gazebo Harmonic

This dev container uses Gazebo Harmonic, the latest long-term support (LTS) release of the Gazebo simulator, providing:

Advanced physics: Realistic dynamics simulation for accurate robot behavior
GPU acceleration: Hardware-accelerated rendering for smooth visualization
Sensor simulation: LiDAR, cameras, IMU, and other sensors with realistic noise models
ROS2 integration: Seamless communication between Gazebo and ROS2 nodes

Gazebo Harmonic replaces the older “Gazebo Classic” and provides improved performance, better ROS2 integration, and modern rendering capabilities.

Accessing Gazebo

Browser-based GUI

Access the Gazebo interface through noVNC:

Start a simulation (e.g., tb3_empty)
Open your browser to http://localhost:6080
Enter password: ros
Gazebo window appears on the virtual desktop

Command line interface

Interact with Gazebo using the gz command:

# List running simulations
gz sim list

# Get world information
gz world info

# Spawn models programmatically
gz model --spawn-file model.sdf

Working with simulations

Launching worlds

Empty world

tb3_empty

Minimal environment for basic testing

TurtleBot3 world

tb3_world

Obstacle course for navigation

House world

tb3_house

Multi-room indoor environment

Simulation features

Physics engine

Gazebo simulates:

Rigid body dynamics: Accurate mass, inertia, and collision detection
Friction models: Realistic wheel-ground interaction
Gravity and forces: External forces affecting robot motion

Sensor models

Available sensors include:

LiDAR: 360-degree laser range finder with configurable resolution
RGB cameras: Visual sensors for computer vision tasks
IMU: Accelerometer and gyroscope for orientation estimation
Odometry: Wheel encoder simulation for dead reckoning

Performance optimization

If Gazebo runs slowly, try these optimization tips from the troubleshooting guide.

GPU acceleration

Verify GPU support:

glxinfo | grep "direct rendering"

If GPU acceleration is unavailable, use software rendering:

export LIBGL_ALWAYS_SOFTWARE=1
tb3_empty

Resource allocation

Allocate more resources in Docker Desktop:

CPUs: 4-6 cores recommended
Memory: 8-16 GB recommended
Disk: 50 GB+ for simulations and logs

Troubleshooting

Common Gazebo issues and solutions:

Gazebo won’t start

Kill existing processes:

pkill -9 gz
pkill -9 gzserver
pkill -9 gzclient

Robot model missing

Verify environment variable:

echo $TURTLEBOT3_MODEL
# Should output: burger

Simulation crashes

Clear Gazebo cache:

rm -rf ~/.gz/

Check logs for detailed error messages:

cat ~/.gz/gazebo/server.log
cat ~/.gz/gazebo/client.log

Learning

Contributing

Gazebo documentation

Official documentation