2D LiDAR SLAM

Use QCOM Cartographer for 2D LiDAR SLAM in a Gazebo simulation. Control mapping, freeze, save, map loading, and relocalization through Cartographer service commands.

Prerequisites

• QIR SDK installed
• Display connected (GUI mode)

Build from Source

radxa@airbox$

git clone https://github.com/qualcomm-qrb-ros/qrb_ros_simulation.git
cd qrb_ros_simulation
source /opt/ros/jazzy/setup.bash
colcon build

Pipeline

office mini Gazebo AMR
  → /scan + /odom + /tf
  → cartographer_node + cartographer_occupancy_grid_node
  → /qrb_slam_command (mapping / freeze / save / load / relocalize)
  → /map + /tracked_pose + /scan_matched_points2 + /submap_list

Run

Use the same environment in all terminals:

radxa@airbox$

source /opt/ros/jazzy/setup.bash
export ROS_DOMAIN_ID=124
export ROS_LOCALHOST_ONLY=1
source qrb_ros_simulation/install/setup.bash

Terminal 1: Start Gazebo Office Scene

Run from the desktop terminal:

radxa@airbox$

ros2 launch qrb_ros_sim_gazebo gazebo_robot_base_mini.launch.py \
  world_model:=office \
  initial_x:=1.0 \
  initial_y:=6.0 \
  enable_depth_camera:=false

Gazebo starts paused. Click the play button at the bottom-left to start the simulation clock.

Terminal 2: Start Cartographer

radxa@airbox$

ros2 launch cartographer_ros qrb_2d_lidar_slam.launch.py use_sim_time:=true

Terminal 3: Interactive Service Client

radxa@airbox$

sh /opt/ros/jazzy/share/cartographer_ros/scripts/qrb_slam_service_request.sh

Select service commands by number:

Command	Meaning	Description
0	Start mapping	Consume /scan and /odom to build a map
1	Finish mapping	Freeze the active mapping trajectory
2	Save map	Save pbstream and export ROS occupancy map
3	Load map	Load the saved map
4	Start localization	Start a localization trajectory against the map
5	Validate relocalization	Check whether localization is ready

Full Workflow

1. Enter 0 in Terminal 3 to start mapping
2. Drive the robot to cover enough static geometry in the environment
3. Enter 1 to freeze the trajectory, then 2 to save the map
4. The map is saved to /var/tmp/cartographer_ros/map/map_0.pbstream
5. Restart Cartographer (press Ctrl+C in Terminal 2 and re-run the launch)
6. Enter 3 → 4 in Terminal 3 to load the map and start localization
7. Move the robot to collect enough scans
8. Enter 5 — expected response: Relocalization Successful, Localization ID: 1

Driving the Robot

Cartographer services control mapping and localization state only; they do not drive the robot. An independent source must publish /cmd_vel during mapping and localization. Always follow any velocity command with a zero Twist to stop:

radxa@airbox$

# Move forward
ros2 topic pub --rate 10 --times 50 /cmd_vel geometry_msgs/msg/Twist \
  '{linear: {x: 0.4}, angular: {z: 0.0}}'
# Stop (Gazebo retains the last velocity command)
ros2 topic pub --times 1 /cmd_vel geometry_msgs/msg/Twist '{}'

Expected Output

Normal responses for commands 0 through 5:

0: Success. Mapping is Started. Trajectory ID: 0
1: TrajectoryID:0 has finished.
2: Write state to pbstream successfully.
3: Load Map Successfully.
4: Success. Localization is Started. Trajectory ID: 1
5: Relocalization Successful, Localization ID: 1

If the mapping route does not cover enough area, command 5 may return Relocalization Not Ready. — more scan data is needed to connect the localization trajectory to the saved map.

Limitations

• Validation used simulated Gazebo LiDAR and odometry; no physical RPLIDAR or robot base was tested.
• RViz visualization and GPU-specific Cartographer features were not verified.
• Migration to real hardware requires compatible /odom and /scan topics.