Map Builder

This package provides RTAB-Map SLAM functionality that can subscribe to camera data from a remote OAK-D driver over ROS2 network.

Features

• Real-time SLAM mapping
• Loop closure detection
• Map visualization
• Subscribes to remote camera topics
• Database persistence

Usage

On the laptop (for mapping):

# Build the package
colcon build --packages-select map_builder

# Source the workspace
source install/setup.bash

# Launch the map builder (subscribes to remote camera)
ros2 launch map_builder map_builder.launch.py robot_name:=oak

Parameters

• robot_name: Name of the remote robot/camera to subscribe to (default: "oak")
• use_sim_time: Whether to use simulation time (default: "false")
• config_file: Path to RTAB-Map configuration file

Subscribed Topics

• /{robot_name}/rgb/image_raw: RGB camera images from robot
• /{robot_name}/rgb/camera_info: RGB camera info from robot
• /{robot_name}/stereo/image_raw: Depth images from robot
• /{robot_name}/vio/odometry: Visual-inertial odometry from robot

Published Topics

• /rtabmap/mapData: Map data
• /rtabmap/grid_map: Occupancy grid map
• /rtabmap/cloud_map: Point cloud map

Network Setup

1. Ensure both robot and laptop are on the same network
2. Set the same ROS_DOMAIN_ID on both machines:

   export ROS_DOMAIN_ID=42

Workflow

1. On the robot: Launch the oakd_driver to publish camera data

   ros2 launch oakd_driver driver.launch.py

2. On the laptop: Launch the map_builder to receive data and build maps

   ros2 launch map_builder map_builder.launch.py

3. Check topics: Verify the data flow between robot and laptop

   ros2 topic list
   ros2 topic echo /oak/rgb/image_raw --max-count 1

Configuration

Edit config/rtabmap_config.yaml to modify SLAM parameters:

• Detection rates
• Feature extraction settings
• Loop closure parameters
• Memory management

Database

Maps are automatically saved to ~/.ros/rtabmap.db. You can visualize saved maps using:

rtabmap-databaseViewer ~/.ros/rtabmap.db