rtab-map Algorithm Mapping

1. Course Content

• Run the example program, using the LiDAR and depth camera to implement rtab-map algorithm mapping.

2. rtab-map Introduction

rtab-map (Real-Time Appearance-Based Mapping) is a vision-based real-time simultaneous localization and mapping (SLAM) algorithm, mainly used for mapping and navigation tasks in fields such as robotics and augmented reality. RTAB-Map has the following characteristics.

• Appearance-based loop closure detection: mainly uses visual features for place recognition

• Real-time performance: ensures real-time performance through a memory management mechanism

• Multi-sensor support: supports RGB-D cameras, stereo cameras, and monocular cameras

• 3D map construction: able to create dense 3D point cloud maps

rtab-map Working Principle

• Front-end processing

  • Acquire image data from sensors

  • Extract visual features (such as SIFT, SURF, ORB, etc.)

  • Calculate the pose transform between the current frame and previous frames

• Back-end optimization

  • Use graph optimization techniques (such as g2o) to optimize the pose graph

  • Detect loop closures (when the robot returns to a previously visited place)

• Memory management

  • Adopt the mechanism of working memory (WM) and long-term memory (LTM)

  • Recent data is kept in WM, and older data is moved to LTM

  • When a loop closure is detected, the relevant data is recalled from LTM back to WM

3. Preparation

ros2 launch yahboomcar_bringup rtab_map.launch.py

4. Case Demonstration

• Control the car's movement; you can use keyboard control or handle control; here keyboard control is taken as an example

xxxxxxxxxx
ros2 run common_demo keyboard_control

• Press I, <, J, and L to control the car's forward, backward, left turn, and right turn respectively; control the car to move slowly to complete the mapping, and press k to stop moving

• After mapping is complete, press ctrl+c to close the mapping program; the map will be automatically saved to the path ~/map/rtabmap.db.

5. Source Code Analysis

• Launch file path:

xxxxxxxxxx
~/yahboomcar_ros2_ws/yahboomcar_ws/src/yahboomcar_bringup/launch/rtab_map.launch.py

• delay_start_rtabmap_slam: started by TimerAction with a 5-second delay, activated only when mode == 'slam', waiting for the other system nodes to be ready before starting mapping

• The first IncludeLaunchDescription: includes the rtabmap mapping launch file, with the key parameters:

  • rgb_topic / depth_topic / camera_info_topic: camera RGB, depth, and intrinsic parameters topics

  • frame_id: robot base coordinate frame base_footprint

  • odom_topic: odometry topic /odom

  • subscribe_scan: subscribe to LiDAR scan data, fusing LiDAR and vision for mapping

  • visual_odometry / icp_odometry: both set to false, using external odometry rather than rtabmap's internal visual odometry

  • approx_sync: enable approximate time synchronization; approx_sync_max_interval sets the maximum time difference to 0.01s

  • sync_queue_size / topic_queue_size: synchronization queue and topic queue size 50

  • rtabmap_args: --delete_db_on_start: clear the old database at each startup and re-map

  • database_path: map save path ~/map/rtabmap.db

  • rviz / rviz_cfg: enable Rviz and load the configuration file

• The second IncludeLaunchDescription: includes car_base.launch.py, enabling the IMU, LiDAR, and camera, disabling the magnetometer, and using EKF odometry as the TF source

xxxxxxxxxx
    delay_start_rtabmap_slam= TimerAction(
        condition=IfCondition(PythonExpression(["'", mode, "' == 'slam'" ])),
        period=5.0,
        actions=[
            IncludeLaunchDescription(
                    PythonLaunchDescriptionSource(rtabmap_launch),
                    launch_arguments={
                        "rgb_topic":                    rgb_topic,
                        "depth_topic":                  depth_topic,
                        "camera_info_topic":            camera_info_topic,
                        "frame_id":                     'base_footprint',
                        "odom_topic":                   '/odom',
                        "use_sim_time":                 'false',
                        "rviz":                         'true',
                        "rviz_cfg":                      rviz_cfg,
                        "database_path":                 database_path,     
                        "rtabmap_viz":                  'false',
                        "approx_sync":                  'true',
                        "approx_sync_max_interval":     '0.01',
                        "qos":                          '0',
                        "visual_odometry":              'false',
                        "icp_odometry":                 'false',
                        "subscribe_scan":               'true',
                        "sync_queue_size":              '50',
                        "topic_queue_size":             '50',
                        "rtabmap_args":                 '--delete_db_on_start',
                        "log_level":                    'info'
                    }.items(),
                ),
            IncludeLaunchDescription(
                PythonLaunchDescriptionSource(os.path.join(get_package_share_directory('yahboomcar_bringup'),'launch','car_base.launch.py')),
                launch_arguments={'log_level': 'info',
                                    'use_imu': 'True',
                                    'use_lidar': 'True',
                                    'use_mag': 'False',
                                    'use_camera': 'True',
                                    'odom_tf_source': 'ekf'
                                  }.items())
        ]
     )