Transform data too old when converting from odom to map

[nedo99]

Bug report

Required Info:

• Operating System:
  • Ubuntu 20.04
• ROS2 Version:
  • Foxy
• Version or commit hash:
  • foxy-devel 2590238de036513c96c6fd2c76884c7599b72a01

Steps to reproduce issue

TURTLEBOT3_MODEL=waffle ros2 launch turtlebot3_gazebo turtlebot3_world.launch.py
ros2 launch slam_toolbox online_async_launch.py
ros2 launch nav2_bringup navigation_launch.py
rviz2

Expected behavior

I am able to set as many 2DNavGoals from rviz2 as I want, even if I set the goal on an unreachable location. I would expect even after setting a goal to an unreachable location to get navigation failed, and afterward, when setting a reachable goal, navigation works as usual.

Actual behavior

After setting >5-6 goals (not at once, one after another one, after it reaches the goal), navigation goes to recovery mode and does spin, wait running, until it aborts. After that, it is not possible to reach even the near goals. It outputs this message most of the time:

[controller_server-6] [INFO] [1619168751.603520587] [controller_server]: Passing new path to controller.
[controller_server-6] [ERROR] [1619168751.603785501] [tf_help]: Transform data too old when converting from odom to map
[controller_server-6] [ERROR] [1619168751.603807308] [tf_help]: Data time: 233s 104000000ns, Transform time: 232s 56000000ns
[controller_server-6] [ERROR] [1619168751.603859730] [controller_server]: Unable to transform robot pose into global plan's frame
[controller_server-6] [WARN] [1619168751.603930573] [controller_server_rclcpp_node]: [follow_path] [ActionServer] Aborting handle.
[controller_server-6] [INFO] [1619168751.618962554] [local_costmap.local_costmap]: Received request to clear entirely the local_costmap
[planner_server-7] [INFO] [1619168751.619290879] [global_costmap.global_costmap]: Received request to clear entirely the global_costmap
[recoveries_server-8] [INFO] [1619507274.773342970] [recoveries_server]: Attempting spin
[recoveries_server-8] [INFO] [1619507274.778353503] [recoveries_server]: Turning 1.57 for spin recovery.
[recoveries_server-8] [INFO] [1619507275.780243732] [recoveries_server]: spin running...

and it goes to spin and waiting after this.

I increased the transform_tolerance to 0.7, but does not matter of the value the message is still there.

Additional information

I am aware of this issue https://github.com/ros2/ros2/issues/1074 and this one https://github.com/ros-planning/navigation2/issues/1652, and that is why added this patch https://github.com/ros-planning/navigation2/compare/main...KavenYau:temporarily-fix-bt-action-node-endless-loop and part of this one https://github.com/ros-planning/navigation2/commit/39ccd457653406bf31f0d59e8084f125ed72f04b (only nav2_behavior_tree/src/behavior_tree_engine.cpp part of the patch, since nav2_behavior_tree/include/nav2_behavior_tree/bt_action_server_impl.hpp is not available in foxy-devel).

Should I reduce controller frequency? I checked system performance and it is constant all the time, so lack of resources should not cause this issue. Did I apply the wrong patches for mentioned issues?

Looking forward to your replies.

Regards,
Nedim

[SteveMacenski]

None of those links are relevant, you have a timing issue in your system and TF is not synchronized well. This is an issue I nor anyone can really help with very much. Your best bet is to start with a working known configuration (like the nav2_bringup turtlebot simulation launch files) and change things 1 thing at a time until you find where your issue lies.

I don't think this is likely an issue in nav2 at all (but I can't say that with absolute certainty), I think it might just be an issue with settings / configurations / etc. At a high level, what you're launching makes sense, but we also support all of that in a single launch file in the nav2_bringup package. I'd try there and see if you can reproduce and then if you can't, close this issue but figure out where you're missing something. Its likely something in a configuration file or related to sim time or available platform resources

[nedo99]

OK, but if there is a config issue, shouldn’t be reported right away and not after setting many goals?

[SteveMacenski]

I can't answer you that, any TF timing issues are so user specific that I can't really help. If you launch our launch file and things work, then its something in your config

[nedo99]

OK. Thanks!

[a-strong-python]

Bug report

Required Info:

• Operating System:

  • Ubuntu 20.04

• ROS2 Version:

  • Foxy

• Version or commit hash:

  • foxy-devel 2590238de036513c96c6fd2c76884c7599b72a01

Steps to reproduce issue

TURTLEBOT3_MODEL=waffle ros2 launch turtlebot3_gazebo turtlebot3_world.launch.py
ros2 launch slam_toolbox online_async_launch.py
ros2 launch nav2_bringup navigation_launch.py
rviz2

Expected behavior

I am able to set as many 2DNavGoals from rviz2 as I want, even if I set the goal on an unreachable location. I would expect even after setting a goal to an unreachable location to get navigation failed, and afterward, when setting a reachable goal, navigation works as usual.

Actual behavior

After setting >5-6 goals (not at once, one after another one, after it reaches the goal), navigation goes to recovery mode and does spin, wait running, until it aborts. After that, it is not possible to reach even the near goals. It outputs this message most of the time:

[controller_server-6] [INFO] [1619168751.603520587] [controller_server]: Passing new path to controller.
[controller_server-6] [ERROR] [1619168751.603785501] [tf_help]: Transform data too old when converting from odom to map
[controller_server-6] [ERROR] [1619168751.603807308] [tf_help]: Data time: 233s 104000000ns, Transform time: 232s 56000000ns
[controller_server-6] [ERROR] [1619168751.603859730] [controller_server]: Unable to transform robot pose into global plan's frame
[controller_server-6] [WARN] [1619168751.603930573] [controller_server_rclcpp_node]: [follow_path] [ActionServer] Aborting handle.
[controller_server-6] [INFO] [1619168751.618962554] [local_costmap.local_costmap]: Received request to clear entirely the local_costmap
[planner_server-7] [INFO] [1619168751.619290879] [global_costmap.global_costmap]: Received request to clear entirely the global_costmap
[recoveries_server-8] [INFO] [1619507274.773342970] [recoveries_server]: Attempting spin
[recoveries_server-8] [INFO] [1619507274.778353503] [recoveries_server]: Turning 1.57 for spin recovery.
[recoveries_server-8] [INFO] [1619507275.780243732] [recoveries_server]: spin running...

and it goes to spin and waiting after this.

I increased the transform_tolerance to 0.7, but does not matter of the value the message is still there.

Additional information

I am aware of this issue https://github.com/ros2/ros2/issues/1074 and this one https://github.com/ros-planning/navigation2/issues/1652, and that is why added this patch https://github.com/ros-planning/navigation2/compare/main...KavenYau:temporarily-fix-bt-action-node-endless-loop and part of this one https://github.com/ros-planning/navigation2/commit/39ccd457653406bf31f0d59e8084f125ed72f04b (only nav2_behavior_tree/src/behavior_tree_engine.cpp part of the patch, since nav2_behavior_tree/include/nav2_behavior_tree/bt_action_server_impl.hpp is not available in foxy-devel).

Should I reduce controller frequency? I checked system performance and it is constant all the time, so lack of resources should not cause this issue. Did I apply the wrong patches for mentioned issues?

Looking forward to your replies.

Regards, Nedim

Have you solved this issue? I'm encountering the same problem. My map->odom->base_link frequency is all above 100Hz, but this issue still occurs.