UniversalRobots / Universal_Robots_ROS_Driver

Universal Robots ROS driver supporting CB3 and e-Series
Apache License 2.0
747 stars 401 forks source link

TF_REPEATED_DATA ignoring data with redundant timestamp for frame tool0_controller at time xxxx according to authority ur_hardware_interface #586

Closed cambel closed 1 year ago

cambel commented 1 year ago

Summary

We have a dual arm set up with two UR5e robots. Since migrating our system to ROS Noetic we experience a spam of warnings coming from ur_hardware_interface.

Versions

Impact

We have a compliance controller streaming motion commands. The performance significantly decreases when the spam of warnings are printed.

Steps to Reproduce

Here is our launch file: https://github.com/o2ac/o2ac-ur/blob/noetic-devel/catkin_ws/src/o2ac_scene_description/launch/osx_arms_bringup.launch

Expected Behavior

No spam of warning messages.

Actual Behavior

The following is a sample of the messages that are spammed in the terminal.

[ WARN] [1668669563.967702653]: TF_REPEATED_DATA ignoring data with redundant timestamp for frame tool0_controller at time 1668669563.965094 according to authority /a_bot/ur_hardware_interface
[ WARN] [1668669564.233413326]: TF_REPEATED_DATA ignoring data with redundant timestamp for frame tool0_controller at time 1668669564.231068 according to authority /b_bot/ur_hardware_interface

In our case the namespace for the robots are a_bot and b_bot.

The spam of warning messages is not continuous but it occurs very frequently even when there is no command sent to the robots.

gavanderhoorn commented 1 year ago

This is https://github.com/ros/geometry2/issues/467.

Are the TF frames prefixed?

Could also be the broadcaster is actually publishing the same transform multiple times. If that's the case, that should be fixed.

cambel commented 1 year ago

Hi @gavanderhoorn

My bad, for lack of use of the tool0_controller, we didn't set up the parameter tf_prefix before. Adding a different tf_prefix to each robot seems to fix the problem.