From time to time the laserscans drift towards the robot in a weird manner, even if robot is not moving.
only in gazebo!

more information following

It seems you are doing good work on porting these packages. I saw that @mintar (Martin Günther) is also working on a ros2 branch. He has been merging some of your commits, but have also contributed a good deal himself. Are you aware of his work, could it make sense to collaborate?

Are you planning to also include MiR 250 in the repository?
I'm new to ROS and unable to replace the MiR 100 with the MiR 250 for the simulation and actual run.

The internal robot PC's is not synchronized (for example via NTP), so it tends to get out of sync quickly (about 1 second per day!). This causes TF transforms timing out etc.

Instead of setting time manually (using the mir.com dashboard) a simple script should be implemented, that synchronizes the robot time with system time.

Since ROS1 remaps mobile_base_controller/cmd_vel and odom to the prefix namespace anyway, the namespace should be removed from the robot_description in the ROS2 pkg.

As the referenced fork (DFKI-NI/rospy_message_converter#56) has been reviewed and the rospy_message_converter fully ported to ROS 2 (https://github.com/uos/rospy_message_converter/tree/galactic) the resource can now be changed to the original repo.

Once it's released as a debian pkg, it can also be excluded from ros2.repos.

I fast too fast to close #1, using the cylinder collision prevented MiR from tilting but I realized the wheels sink into the floor.

Quick check with the tf showed that the collision needs to be rotated (mesh already had that rotation included). The screenshot shows the cylindric collision as the visual.

• Missing collision mesh for wheels in collision folder
  To speed up simulation, mesh needs to be scaled down and replaced in robot urdf here and here

• Check existing base mesh as well and improvement simulation when scaled down more

Executables of mir_robot package are not installed properly, even though the setup.cfg files look fine.

Steps to reproduce:

1. purge workspace
2. build mir_robot package
3. source setup file
4. TRY ros2 launch mir_driver mir_headless_launch.py OR
5. TRY ros2 run mir_driver mir_bridge

Expected behaviour:

• the driver starts

Observed behaviour:

Launch/Run crashes with the following error

    launch.substitutions.substitution_failure.SubstitutionFailure: package 'mir_driver' found at '/home/soenke/dev/tend-o-bot/install', but libexec directory '/home/soenke/dev/tend-o-bot/install/lib/mir_driver' does not exist

Any ideas on why this happens? I have completely reinstalled my ROS galactic distribution but the error still persists.

The laserscan merger works with a couple of params to determine how the scans should be merged & published.

The params set for the driver work as expected and a solid laserscan is the result. In simulation the params need a review since there is a flickering of the scan happening, which not only confuses the user but also has a great impact on the localization and navigation process.

The usage of a namespace in mir_description and mir_gazebo is almost fully supported.

• Add namespace parser functionality for laserscan_multi_merger topics
• Add namespace for mir_navigation package

more information following

Noticed non-realistic movements in gazbeo simulation of the wheels when forces are applied.

Steps to reproduce:

1. Apply any kind of force on robot base.

or

2. Use teleop twist to spin the robot and stop velocity commands. Robot is drifting further due to momentum > friction.

• Collision friction and joint damping should be checked and tuned.
• Check inertia values as well