Skip to main content
SLAM (Simultaneous Localization and Mapping) enables the robot to build a map of an unknown environment while simultaneously tracking its location within that map. The TurtleBot3 uses Cartographer for SLAM functionality.

​
What is SLAM?

SLAM algorithms process sensor data (laser scans, odometry) to:
  • Localize: Determine the robot’s position and orientation
  • Map: Build a 2D occupancy grid of the environment
  • Update: Continuously refine both the map and position estimate
The result is an accurate map that can be used for autonomous navigation.

​
Starting SLAM

​
Using the alias

The quickest way to start SLAM with Cartographer: 
tb3_slam
This launches Cartographer with simulation time enabled and automatically opens RViz2 for visualization.

​
Using the full command

ros2 launch turtlebot3_cartographer cartographer.launch.py use_sim_time:=True
The use_sim_time:=True parameter is critical for simulation. It tells ROS2 to use Gazebo’s simulated time instead of wall clock time.

​
Complete SLAM workflow

1

Launch simulation world

Start a Gazebo world with obstacles to map:
tb3_world
Wait for the simulation to fully load.
2

Start SLAM

In a second terminal, launch Cartographer:
tb3_slam
RViz2 will open automatically in the VNC desktop.
3

Start teleoperation

In a third terminal, start keyboard control:
tb3_teleop
4

Drive to explore

Use keyboard controls to drive the robot:
  • Press i to move forward
  • Use j and l to rotate
  • Drive slowly through all areas
  • Ensure complete environment coverage
5

Monitor in RViz2

Access the VNC desktop at http://localhost:6080 and watch:
  • The map being built in real-time (gray = unknown, white = free space, black = obstacles)
  • The robot’s position and laser scan data
  • The particle cloud showing position uncertainty
6

Save the map

Once mapping is complete, save the map in a fourth terminal:
mkdir -p ~/maps
ros2 run nav2_map_server map_saver_cli -f ~/maps/my_map
This creates two files:
  • my_map.pgm (map image)
  • my_map.yaml (map metadata)

​
Mapping tips

​
Movement strategy

Move slowly: Rapid movements reduce map quality
  • Use brief taps of movement keys instead of holding
  • Allow the SLAM algorithm time to process sensor data
  • Pause occasionally at key locations
Smooth turns: Gradual rotations work better than sharp spins
  • Use u and o keys for curved paths
  • Avoid rapid rotation with j and l
Complete coverage: Drive through all areas you want mapped
  • Get close to walls and obstacles for detail
  • Cover open spaces thoroughly
  • Don’t skip any areas
Close the loop: Return to your starting position
  • This helps the SLAM algorithm reduce drift
  • Improves overall map accuracy
  • Essential for large environments

​
Choosing the right world

For practice: Start with tb3_world
  • Medium complexity
  • Clear obstacles
  • Good for learning SLAM basics
For complex mapping: Use tb3_house
  • Multiple rooms
  • Realistic indoor environment
  • Tests advanced SLAM capabilities
Avoid: Don’t use tb3_empty for SLAM practice
  • No features to map
  • SLAM needs distinct landmarks
  • Use only for testing SLAM startup

​
Understanding RViz2 visualization

When you launch SLAM, RViz2 displays several important elements:

​
Map display

  • White areas: Free space where the robot can travel
  • Black areas: Obstacles and walls detected by sensors
  • Gray areas: Unknown/unexplored regions
  • Map updates: Watch gray areas turn white or black as you explore

​
Robot position

  • Red arrow: Current robot pose (position and orientation)
  • Green/blue particles: Position estimate uncertainty
  • Laser scan: Red dots showing current sensor readings

​
Quality indicators

  • Clear boundaries: Sharp transitions indicate good map quality
  • Fuzzy edges: Blurry boundaries suggest you moved too fast
  • Gaps in walls: Missing data means you need to explore that area more

​
Saving your map

Once you’ve finished exploring and the map looks complete, save it for future use.

​
Create maps directory

mkdir -p ~/maps

​
Save the map

ros2 run nav2_map_server map_saver_cli -f ~/maps/my_map
Use descriptive names for your maps based on the environment:
ros2 run nav2_map_server map_saver_cli -f ~/maps/turtlebot3_world
ros2 run nav2_map_server map_saver_cli -f ~/maps/house_floor1

​
Generated files

The map saver creates two files: my_map.pgm
  • Portable graymap image file
  • Visual representation of the map
  • Can be viewed with image viewers
my_map.yaml
  • Map metadata and configuration
  • Contains resolution, origin, and thresholds
  • Required by Navigation2
Both files are needed for navigation. Keep them together and don’t rename them separately.

​
Workflow examples

​
Mapping TurtleBot3 world

Terminal 1: 
tb3_world
Terminal 2: 
tb3_slam
Terminal 3: 
tb3_teleop
# Drive slowly through the entire world
# Cover all areas around obstacles
# Return to starting position
Terminal 4: 
mkdir -p ~/maps
ros2 run nav2_map_server map_saver_cli -f ~/maps/turtlebot3_world

​
Mapping house environment

Terminal 1: 
tb3_house
Terminal 2: 
tb3_slam
Terminal 3: 
tb3_teleop
# Explore each room systematically
# Drive through all doorways
# Map furniture and obstacles
# Return to start to close the loop
Terminal 4: 
ros2 run nav2_map_server map_saver_cli -f ~/maps/house_complete

​
Troubleshooting

​
RViz2 doesn’t open

1

Check VNC access

Ensure you can access the desktop at:
http://localhost:6080
2

Manually launch RViz2

In the VNC desktop, open a terminal and run:
rviz2
3

Load SLAM config

In RViz2, load the appropriate configuration file for SLAM visualization.

​
Map quality is poor

Move slower
  • Reduce linear velocity by pressing x several times
  • Use brief key taps instead of holding keys down
Smoother movements
  • Avoid sharp rotations
  • Use curved paths with u and o keys
  • Don’t make sudden direction changes
Better coverage
  • Drive closer to walls and obstacles
  • Explore all areas thoroughly
  • Return to starting position to close the loop

​
No map appears in RViz2

1

Verify SLAM is running

Check that Cartographer nodes are active:
ros2 node list | grep cartographer
2

Check topics

Verify the map topic is publishing:
ros2 topic list | grep map
ros2 topic echo /map --once
3

Source workspace

Ensure the workspace is properly sourced:
sb

​
Map save fails

1

Create directory

Ensure the maps directory exists:
mkdir -p ~/maps
2

Check permissions

Verify write permissions:
ls -ld ~/maps
3

Verify map topic

Confirm the map is being published:
ros2 topic hz /map
Should show a steady update rate.

​
Advanced SLAM options

​
Different SLAM algorithms

While this environment uses Cartographer by default, TurtleBot3 also supports:
  • SLAM Toolbox
  • Gmapping (ROS1, requires bridging)

​
Parameter tuning

For advanced users, Cartographer parameters can be adjusted in the launch file to optimize for:
  • Different sensor configurations
  • Varying environment types
  • Speed vs. accuracy trade-offs

​
Real-time map quality

Monitor these indicators in RViz2:
  • Particle spread (localization confidence)
  • Loop closure events (drift correction)
  • Edge clarity (measurement quality)

Build docs developers (and LLMs) love

Get started for freeTalk to us