Added nav2 simple commander with the new Gazebo

[ahcorde]

Basic Info

Info	Please fill out this column
Ticket(s) this addresses	#2997

---

Description of contribution in a few bullet points

I created a new package called nav2_gz_simple_commander to replicate the package nav2_simple_commander but this time with the new Gazebo.

I have some issues that maybe you can help me to solve. The odom frame should be locate close to the map frame (compared with the gazebo-classic simulation), but for some reason is located where the robots is spawned and sometimes the robot is lost. You can check this in the gif. any thoughts?

NOTE: You need to add this repository to your workspace https://github.com/ahcorde/aws-robomaker-small-warehouse-world

Test it

ros2 launch nav2_gz_simple_commander nav_to_pose_example_launch.py

[SteveMacenski]

The only differences are the launch files, no? If so, delete this package and just add the complimentary launch files to the existing package. I don't think there's any reason to have an entirely separate package.

[ahcorde]

The only differences are the launch files, no? If so, delete this package and just add the complimentary launch files to the existing package. I don't think there's any reason to have an entirely separate package.

Done here 6e708e9

Any thoughts about why the robot is lost ?

[SteveMacenski]

Probably some of the URDF things I pointed out in the other PR. A 5m lidar is completely insufficient to be localizing in a warehouse of that size.

Otherwise, your changes look good if you make the updates for the rest of these, but I'd recommend just waiting until we get the other PR done since this is blocked by that anyway

[ahcorde]

Probably some of the URDF things I pointed out in the other PR. A 5m lidar is completely insufficient to be localizing in a warehouse of that size.

Otherwise, your changes look good if you make the updates for the rest of these, but I'd recommend just waiting until we get the other PR done since this is blocked by that anyway

I noticed that in this case the warehouse.world file contains a different instance of the turtlebot3 which is defined inside the file. In particular the model working on Gazebo classic has a 3.5m lidar.

the only different that I have between models are the type caster wheel, one is ball and in the other one if fixed.

[SteveMacenski]

I can't answer that for you then, maybe the Ignition model you selected has problems?

[SteveMacenski]

Figure out the reason and/or run each of these launch files to verify they work?

This PR I'll broadly just give a once over and merge as long as someone's run them and the demos still work

[ahcorde]

Figure out the reason and/or run each of these launch files to verify they work?

This PR I'll broadly just give a once over and merge as long as someone's run them and the demos still work

I checked all the new launch files and they have the same behaviour as Gazebo classic.

[SteveMacenski]

https://app.circleci.com/pipelines/github/ros-planning/navigation2/9552/workflows/684b5f23-95df-44ff-85f6-20c9897cc0e9/jobs/30626/tests

Some linting problems

[SteveMacenski]

Approved beyond the minor linting problems needed for CI to not have errors

[ahcorde]

humm the CI failures look odd, I can see two tests failing in nav2_gz_simple_commander.test.test_copyright but this package was removed, same happens with nav2_gz_simple_commander.test.test_flake8.

should I need to fix linters that not included in the files that I included ?

[SteveMacenski]

Might be a caching issue, change all 3 instances of v14 to v15 in this file and push it: https://github.com/ros-planning/navigation2/blob/main/.circleci/config.yml

[SteveMacenski]

That worked

[SteveMacenski]

How does this URDF differ from that in nav2_bringup once the other PR is merged? is this necessary?

[SteveMacenski]

It just occurred to me that we abstracted out the gz tb3 robot simulation launch file to deal with the gz specifics / bridges and whatnot. Can't that be used in all these launch files as well if we send it the updated starting pose?

[ahcorde]

I made the PRs independent, because the tb3 for Gazebo classic is defined in the world file inside this package. We are using a two versions of the tb3, which are a little bit different between them.

But I'm happy to use the new resources included in nav2_bringup. What fo you prefer ?

• Keep models independent from nav2_bringup
• Use the same models here and in nav2_bringup

[SteveMacenski]

Same - unless there are key differences... which I don't believe there should be? I think when I made this I may have made the lidar for the simple commander range much higher to deal with the larger nature of the space that would need to be kept.

So I'd say I'd rather re-use as much as possible. Where there are changes where they're not the same, err on the side of less restrictive (e.g. longer ranges and such)

[mergify]

This pull request is in conflict. Could you fix it @ahcorde?

[SteveMacenski]

#4417 supersedes based on https://github.com/ros-navigation/nav2_minimal_turtlebot_simulation