Skip to main content
The TurtleBot3 platform comes in three hardware configurations. This container supports all three models for simulation.

​
Available models

The TURTLEBOT3_MODEL environment variable determines which robot configuration is loaded in simulations.

​
Burger

  • Default model in this container
  • Smallest and most affordable TurtleBot3
  • 2-wheel differential drive
  • LiDAR sensor (360° laser scanner)
  • IMU sensor
  • Ideal for learning and basic navigation

​
Waffle

  • Larger platform with more sensors
  • 2-wheel differential drive
  • LiDAR sensor
  • Intel RealSense R200 camera
  • IMU sensor
  • Better for advanced navigation and vision tasks

​
Waffle Pi

  • Similar to Waffle but with Raspberry Pi Camera
  • 2-wheel differential drive
  • LiDAR sensor
  • Raspberry Pi Camera v2
  • IMU sensor
  • Alternative to Waffle with different camera hardware

​
Current model configuration

The default model is set in .devcontainer/devcontainer.json:49-57: 
"remoteEnv": {
  "DISPLAY": ":1",
  "ROS_DOMAIN_ID": "30",
  "TURTLEBOT3_MODEL": "burger",
  "RMW_IMPLEMENTATION": "rmw_cyclonedds_cpp",
  "GZ_VERSION": "harmonic",
  "QT_QPA_PLATFORM": "xcb",
  "LIBGL_ALWAYS_SOFTWARE": "${localEnv:LIBGL_ALWAYS_SOFTWARE:0}"
}

​
Changing the model

​
Method 1: Rebuild container (permanent)

To permanently change the model:
  1. Edit .devcontainer/devcontainer.json
  2. Change the TURTLEBOT3_MODEL value:
"remoteEnv": {
  "TURTLEBOT3_MODEL": "waffle"  // Options: burger, waffle, waffle_pi
}
  1. Rebuild the container:
    • Press F1 in VS Code
    • Select Dev Containers: Rebuild Container
    • Wait for rebuild to complete (5-10 minutes)
Rebuilding the container will stop all running processes and re-run the build scripts. Any unsaved work in terminals will be lost.

​
Method 2: Set per-session (temporary)

To temporarily use a different model in your current terminal session: 
export TURTLEBOT3_MODEL=waffle
Then launch your simulation: 
tb3_empty  # Will now spawn the waffle model
This method only affects the current terminal. Open a new terminal and it will revert to the default model.

​
Method 3: Set per-launch (one-time)

Override the model for a single launch: 
TURTLEBOT3_MODEL=waffle_pi ros2 launch turtlebot3_gazebo empty_world.launch.py

​
Verifying the active model

Check which model is currently configured: 
echo $TURTLEBOT3_MODEL
This should output one of: burger, waffle, or waffle_pi

​
Model differences in simulation

​
Visual differences

  • Burger: Compact, circular base
  • Waffle: Larger, square base with camera mount
  • Waffle Pi: Similar to Waffle with different camera

​
Sensor topic differences

All models publish:
  • /scan - LiDAR data
  • /odom - Odometry
  • /imu - IMU data
  • /cmd_vel - Velocity commands (input)
Waffle and Waffle Pi additionally publish:
  • /camera/image_raw - Camera images
  • /camera/camera_info - Camera calibration

​
Performance differences

The Waffle and Waffle Pi models require slightly more computational resources due to camera simulation.

​
Model files location

Model descriptions are stored in the TurtleBot3 source: 
/workspace/turtlebot3_ws/src/turtlebot3/turtlebot3_description/
├── urdf/
│   ├── turtlebot3_burger.urdf
│   ├── turtlebot3_waffle.urdf
│   └── turtlebot3_waffle_pi.urdf

​
Troubleshooting

​
No robot appears in Gazebo

Symptom: Gazebo launches but the world is empty Solution:
  1. Verify the model is set: 
    echo $TURTLEBOT3_MODEL
  2. If empty, set it: 
    export TURTLEBOT3_MODEL=burger
  3. Relaunch the simulation

​
Wrong model appears

Symptom: Expected Waffle but Burger appeared Solution: The environment variable may not be set in the terminal you’re using. Either:
  • Source your bashrc: source ~/.bashrc
  • Export the model manually: export TURTLEBOT3_MODEL=waffle
  • Restart the terminal

​
Model mismatch errors

Symptom: ROS2 topics don’t match expected model Solution: Make sure all terminals use the same model. Run echo $TURTLEBOT3_MODEL in each terminal and ensure they match.

Build docs developers (and LLMs) love

Get started for freeTalk to us