Publish ground truth robot position as dynamic TF#759
Publish ground truth robot position as dynamic TF#759peci1 wants to merge 11 commits intoosrf:masterfrom
Conversation
|
Build finished. 15 tests run, 0 skipped, 0 failed. |
adlarkin
left a comment
There was a problem hiding this comment.
Thanks for working on this! The changes here look good, but I'm not approving it yet because #745 combines a few of the circuit launch files into one - so, once #745 is merged, I think it would be good to re-base this on master, and then apply the ground truth update to just the newly combined competition.ign file instead of applying it separately to the launch files for cave, tunnel, and urban.
Isn't there really a way to just some common file defining all these common definitions for all robots? This approach seems pretty awkward to me...
This functionality should be available, but I'm not sure that we've taken advantage of it yet (I'd have to double-check). Perhaps we can make the include update in this PR, or in a separate PR after merging this.
|
Okay, I can rebase on top of #745. Or do you expect it to change a lot until merged? And it'd be great to utilize some include facility, I can also add it to this PR. Is there some documentation/examples for it? |
#745 hasn't been reviewed yet, so I'd wait to rebase in case any changes happen there. I don't think much will change, but perhaps reviewers for that PR will catch something that I missed originally. It's probably best to rebase this PR on
I don't think there is any documentation for this, but I have tested a possible solution with
I'm no ruby expert, so if you have a better idea for the |
|
Thanks for the hint, so I'll wait for #745. Isn't there a way to achieve the include via SDF? That would look more natural to me... |
|
And what would you think about extracting more pieces to the common include file - like gas sensor or joint state publisher? |
I don't believe that there is. The So, we will probably need to do something with ruby... but maybe there's a cleaner solution than the one I proposed above.
I think it would be worth using some sort of "include" mechanism for all common parts of model files. If we want to do something like this, it would probably be best to leave it for a separate PR, since this PR is really meant for fixing the topic for the ground truth data. So, after some more thought, I'm thinking that the best approach moving forward would be this:
There are a few things to keep in mind:
What do you think about this approach? Any thoughts/objections? |
# Conflicts: # subt_ign/launch/cave_circuit.ign # subt_ign/launch/tunnel_circuit_practice.ign # subt_ign/launch/urban_circuit.ign
|
I updated this PR with the changes from #745 and also applied the changes to the newly accepted robots, so it's ready to be merged. |
|
Build finished. 21 tests run, 0 skipped, 1 failed. |
There were a few more robot configurations added in the last week or two that might need these updates as well:
Yeah, I'll take a look (sorry for the delayed response). If #786 is based off of this PR, do we still need this PR? Or can we close it? |
# Conflicts: # subt_ign/launch/virtual_stix.ign # subt_ign/launch/virtual_stix_headless.ign
Signed-off-by: Martin Pecka <peckama2@fel.cvut.cz>
|
Updated. I did a search throughout whole repo and all files except validate_connections.ign show an even number of occurences of |
|
Build finished. 21 tests run, 0 skipped, 0 failed. |
|
I've updated this PR with the latest changes from master (no merge conflicts). |
|
I think this PR would still bring value in the simulator. However, it'd need a major update since it was last synced with master in March... Is the idea of this PR still welcome? |
# Conflicts: # submitted_models/ctu_cras_norlab_absolem_sensor_config_1/launch/spawner.rb # submitted_models/ctu_cras_norlab_absolem_sensor_config_2/launch/spawner.rb
|
I've updated this PR so that it handles all models currently in |
Solves #659 and #607. I hope I got all the places where ground truth publishing might be controlled.
Isn't there really a way to just
<include>some common file defining all these common definitions for all robots? This approach seems pretty awkward to me...