Upgrade code to ROS Noetic

[Reschivon]

What is a quick description of the change?

Python2 code, specifically the slam_mode_goal code, has been changed to work in python 3.
The teb_local_planner_tutorial package, which does not appear to be used, has been removed because it is no longer a package in Noetic

Is this fixing an issue?

No

Were any issues created as a result of this change?

Velodyne code uses functions that are now deprecated. Simple fix would be reimporting this code from the Velodyne SDK

Are there more details that are relevant?

Check lists (check x in [ ] of list items)

• Test written/updating
• Tests passing
• Coding style (indentation, etc)

I'm not sure what the standard for test cases are, so there are no test cases

Any additional comments?

Please verify that everything works on your machine.

[alanssitis]

Are we looking to upgrade to Noetic? It will make it easier on on boarding members since newer Ubuntu releases have better support. Also, Python3 is very helpful, and has more documentation.

[Reschivon]

I'm in favor of upgrading. While we might not use ROS 1 for v2, the upgrade is trivial to do and shouldn't have any problems

[alanssitis]

The main issue, however, is that the ROS version on the Jetson would need to be updated, and there may be some recompiling needed. @zghera , thoughts?

[zghera]

I think the main problem would be having to upgrade everyone's VMs and (more importantly) the Jetson to Ubuntu 20.04. Which we could probably do, but it may be more hassle than it's worth. I will ask @ihagedo what he thinks about it since he knows the current state of the Jetson best.

[Reschivon]

Originally I thought this was a good idea because it'd help the new members learn the code better. Is v1 going to be something we finish in a hurry and replace, or will it be something the new members are supposed to build off?

[alanssitis]

To my understanding, v2 will be built from the ground up. Thus making this more of a hassle. However, VMs would need to be updated regardless when switching to v2, so that's probably where one could see a reason to update now rather than later, especially if we stick with ROS 1.

On the other hand, we should not upgrade if we plan to build v2 using ROS 2, it will be a lot of work for very little gain and almost pointless.

[zghera]

In my opinion, v1 should be "archived" once we get v2 development started. The reason is that the use of the Navigation stack really limits our flexibility, so I think shifting our focus to v2 would be the best of resources. I am open to discussing keeping v1 alive if we could use it to train new members in some way.

To @alanssitis point, it would make sense to upgrade now if we knew what version we were going to use in v2. But I have no idea what makes sense for that yet. So unless upgrading to noetic provides an advantage while wrapping up v1 or we decide to extend the life of v1 past the ground demo (we can talk more about this offline), I say we wait on this.

[zghera]

Added the wontfix label so we don't keep getting reminded in Slack. When we determine how hard it is to upgrade the Ubuntu version on the Jetson, @ihagedo or I will update this issue and hopefully approve it as well.

[alanssitis]

In this case, could we possibly do something like this:

1. Get another Jetson
2. Install 20.04 on it
3. Test ASSv1 on Noetic on the Jetson

Even in the case were upgrading ROS distros is off the list, updating OS should be looked at for firmware updates and stability. We can rely on Dockerfiles to run the older version of ROS.

[zghera]

I will ping @ihagedo to try upgrading the OS version on the Jetson and try to install Noetic. If he doesn't feel comfortable doing so or it doesn't work, we will ask Jeremy to order us a new Jetson ASAP.