[develop/fetch][gazebo] Gazebo for noetic

[mqcmd196]

• Build support for gazebo in noetic
• Add launch arg to turn off the google_analyze_text, cpu_monitor

[mqcmd196]

TODO:

• Enable move_base
• Check 73B2 environment
• Check on the real robot whether it doesn't break the function for real robot

[mqcmd196]

@yoshiki-447 Try this

[pazeshun]

• Please update README.
  

  jsk_robot/jsk_fetch_robot/README.md

  Line 47 in 25dd087

  wstool merge -t . https://raw.githubusercontent.com/jsk-ros-pkg/jsk_robot/master/jsk_fetch_robot/jsk_fetch_user.rosinstall.$ROS_DISTRO

  should be updated???
• Please follow [gazebo] Resupport gazebo environment Fetch and PR2 #1912 (comment)

[yoshiki-447]

When I executed the following:

roslaunch fetch_gazebo simulation.launch
roslaunch fetch_moveit_config move_group.launch

rlwrap roseus
(load "package://fetcheus/fetch-interface.l")
(fetch-init)

I faced the following warning periodically on rlwrap roseus:

Warning: TF_REPEATED_DATA ignoring data with redundant timestamp for frame base_link (parent odom) at time 720.438000 according to authority unknown_publisher
         at line 278 in /tmp/binarydeb/ros-noetic-tf2-0.7.7/src/buffer_core.cpp

cc @pazeshun

[yoshiki-447]

To remove this warning, ZebraDevs/robot_controllers#83 is required.

cc @pazeshun

[mqcmd196]

BTW, I can't see any TF_REPEATED_DATA warning in my euslisp console with current jsk_fetch_user.noetic.repos . When does the warning spawn?

[pazeshun]

If you use a high-performance PC, you may not see warnings.
Could you also check in a relatively low-performance PC?

A warning appears when gazebo publishes tf from odom to base_link whose timestamp is the same as tf published previously.
I think this is because ros::Time::now()
https://github.com/ZebraDevs/robot_controllers/blob/938b67673a31d0a495d016e4a31c75263258092f/robot_controllers/src/diff_drive_base.cpp#L375
inside the callback publishing tf at 100Hz
https://github.com/ZebraDevs/robot_controllers/blob/938b67673a31d0a495d016e4a31c75263258092f/robot_controllers/src/diff_drive_base.cpp#L130-L132
sometimes returns the same timestamp as the previous callback.

I assume this problem occurs in relatively low-performance PC where gazebo is slow and ROS Clock may not be updated well.
We use ThinkPad P16s Gen2.
cf. http://wiki.ros.org/roscpp/Overview/Timers#Wall-clock_Timers

[mqcmd196]

Thanks for your info. I'll check it. BTW, I updated some repositories for enabling move_base. Please try with ``` cd catkin_ws/src vcs pull ``` 2024年5月21日(火) 11:28 Shun Hasegawa ***@***.***>:

…

If you use a high-performance PC, you may not see warnings. Could you check this? A warning appears when gazebo publishes tf from odom to base_link whose timestamp is the same as tf published previously. I think this is because ros::Time::now() https://github.com/ZebraDevs/robot_controllers/blob/938b67673a31d0a495d016e4a31c75263258092f/robot_controllers/src/diff_drive_base.cpp#L375 inside the callback publishing tf at 100Hz https://github.com/ZebraDevs/robot_controllers/blob/938b67673a31d0a495d016e4a31c75263258092f/robot_controllers/src/diff_drive_base.cpp#L130-L132 sometimes returns the same timestamp as the previous callback. I assume this problem occurs in relatively low-performance PC where gazebo is slow and ROS Clock <http://wiki.ros.org/Clock> may not be updated well. cf. http://wiki.ros.org/roscpp/Overview/Timers#Wall-clock_Timers — Reply to this email directly, view it on GitHub <#1913 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGUARFHBUP2DQ62SO5TSB3LZDKWM3AVCNFSM6AAAAABH7C5Y72VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMRRGU4TCMZTHE> . You are receiving this because you authored the thread.Message ID: ***@***.***>

-- =========================================== Yoshiki Obinata The University of Tokyo Graduate School of Information Science and Technology Department of Mechano Informatics 2nd grade Ph.D course student JSK Robotics Laboratory ===========================================

[mqcmd196]

I confirmed the TF_REPEATED_DATA spawns on the console. Please unset ZebraDevs/robot_controllers#83 as draft.

[sktometometo]

LGTM