Map does not grow after its initialization

[mac137]

Required Info:

• Operating System:
  • Ubuntu 20.04
• Installation type:
  • I installed this package both via sudo apt install ros-foxy-slam-toolbox as well as from the source (with proper sourcing of the environment). The same behavior in both cases.
• ROS Version
  • ROS2, foxy
• Version or commit hash:
  • 48e97304618ef5b18aa38cf2ea12c41202ebc092, foxy-devel-unfixed
• Laser unit:
  • RP Lidar A1 and others downloaded from the this git repo's issues

Steps to reproduce issue

I provide the static transform between my lidar base_scan and odom frame 'odom' in online_sync_launch.py:

    odom_lidar_static_tf = Node(
        name='tf2_ros_odom_lidar',
        package='tf2_ros',
        executable='static_transform_publisher',
        output='screen',
        arguments=['0', '0', '0', '0.0', '0.0', '0.0', 'odom', 'base_scan'],
    )

    ...

    ld.add_action(odom_lidar_static_tf)

    ...

... and I modify mapper_params_online_sync.yaml like this:

    # ROS Parameters
    odom_frame: odom
    map_frame: map
    base_frame: base_scan
    scan_topic: /scan

and then I do ros2 launch slam_toolbox online_sync_launch.py

Expected behavior

The map should grow and build progressively as my lidar moves.

Actual behavior

The map stays the same after its initialization by the package like this:

and all the time after the initialization, the map stays as above even though the lidar moved and rotated a lot:

Additional information

After running ros2 run tf2_tools view_frames.py I get this TF tree:

and when I check the transform between map and odom is stays the same all the time:

ros2 run tf2_ros tf2_echo map odom
[INFO] [1631990416.761771843] [tf2_echo]: Waiting for transform map ->  odom: Invalid frame ID "map" passed to canTransform argument target_frame - frame does not exist
At time 1631990417.939988347
- Translation: [0.000, 0.000, 0.000]
- Rotation: in Quaternion [0.000, 0.000, 0.000, 1.000]
At time 1631990418.939988140
- Translation: [0.000, 0.000, 0.000]
- Rotation: in Quaternion [0.000, 0.000, 0.000, 1.000]
At time 1631990419.939991397
- Translation: [0.000, 0.000, 0.000]
- Rotation: in Quaternion [0.000, 0.000, 0.000, 1.000]
At time 1631990420.940001147
- Translation: [0.000, 0.000, 0.000]
- Rotation: in Quaternion [0.000, 0.000, 0.000, 1.000]

The console displays this all the time:

ros2 launch slam_toolbox online_sync_launch.py 
[INFO] [launch]: All log files can be found below /home/user_name/.ros/log/2021-09-18-19-52-49-142812-u20-21619
[INFO] [launch]: Default logging verbosity is set to INFO
[INFO] [static_transform_publisher-1]: process started with pid [21621]
[INFO] [sync_slam_toolbox_node-2]: process started with pid [21623]
[INFO] [rviz2-3]: process started with pid [21625]
[static_transform_publisher-1] [INFO] [1631991169.189978256] [tf2_ros_fp_odom]: Spinning until killed publishing transform from 'odom' to 'base_scan'
[sync_slam_toolbox_node-2] [INFO] [1631991169.211974174] [slam_toolbox]: Node using stack size 40000000
[sync_slam_toolbox_node-2] [INFO] [1631991169.223044981] [slam_toolbox]: Using solver plugin solver_plugins::CeresSolver
[sync_slam_toolbox_node-2] [INFO] [1631991169.223135162] [slam_toolbox]: CeresSolver: Using SCHUR_JACOBI preconditioner.
[rviz2-3] [INFO] [1631991169.785774056] [rviz2]: Stereo is NOT SUPPORTED
[rviz2-3] [INFO] [1631991169.785861010] [rviz2]: OpenGl version: 4.6 (GLSL 4.6)
[rviz2-3] [INFO] [1631991169.812876302] [rviz2]: Stereo is NOT SUPPORTED
[sync_slam_toolbox_node-2] Info: clipped range threshold to be within minimum and maximum range!
[sync_slam_toolbox_node-2] Registering sensor: [Custom Described Lidar]
[rviz2-3] [INFO] [1631991173.253112738] [rviz2]: Trying to create a map of size 115 x 79 using 1 swatches

Any idea what I do wrong? Thanks!

[mac137]

Ofc the map does not grow because I put a static transform between odom and base_frame which is wrong. There should be an odometry method that keeps computing the relative pose between these two frames.

[amjack0]

Ofc the map does not grow because I put a static transform between odom and base_frame which is wrong. There should be an odometry method that keeps computing the relative pose between these two frames.

Hi @maciej-3

I am having the similar issue. Did providing the odometry method between two frames worked? How did you solve this?

Regards,

[mac137]

Hi @amjack0

I wrote a simple scan-to-scan odometry node for my lidar and it works good enough for this slam package.