search

Team-Isaac-Polito / reseq_ros2

High-level ROS2 code to control the ReseQ robot
GNU General Public License v3.0
0 stars 0 forks source link

Make agevar compatible with robot and Gazebo #7

Closed GBergatto closed 6 months ago

GBergatto commented 1 year ago

Agevar needs to publish to both diff_cont_X and reseq/moduleX/motors/setpoint

GBergatto commented 1 year ago

@marc0777 We need to rewrite the launch files inside gazebo and then we should be good to go, right?

marc0777 commented 1 year ago

Yes, but to be honest I don't know what to rewrite in particular and how, since the configuration file is hard-coded inside the robot description:

https://github.com/Team-Isaac-Polito/reseq_ros2/blob/4a9232a473ddb054e590d0b1f5a4474bd40532ee/reseq_gazebo/description/reseq_parameters.xacro#L6-L10

This is extremely not portable and doesn't for example even work on my setup, for the usual problem that I my ROS workspace is different than the configuration suggested in the readme.

However I'm afraid that to improve this we would need to therefore change one (or more) of the xacro files, and maybe it's not the right moment for that. What do you think?

Tizio0o0o0o commented 1 year ago

Can I help? I'm not quite sure about the problem, can you explain it?

If the idea is to tell the gazebo model witch config to use, In my opinion this should not be done: the mK1 robot has a different structure and a new gazebo model have to be written. that model will have only mK1 config settings. A good idea might be to modify the gazebo launch file making it able to launch mk1 or mk2 files when requested (when they will exist). Currently the mK1 files do not exist, some problem (like yaw feedback, the agevar script needs these feedback to work) have to been solved (in mK2) after that, creating mK1 files will be easy.

marc0777 commented 1 year ago

The difference in structure is also one of the reason for whch I'd postpone this problem for now, even though we'll have to address it sooner than later (we are already accumulating more backlog than we should, considering we're only one month into this repository).