Robustify OSR launch script, and make rover full stop the wheels when idle

[apollokit]

• made the launch script more robust
  • using symbolic links now so that updates get captured in the systemd service
  • allow usage of custom osr_mod.launch file if you want to disclude e.g. the led screen node
• made the rover truly idle when it's supposed to be idle
  • now we command the wheels to fully stop when a command hasn't been received for a little bit. This consumes less current, and avoids the chattering noise that was present when idle before.

[Achllle]

Haven't looked at this properly but one quick comment: these two changes are sufficiently independent to warrant 2 PRs. That's good practice and helps with debugging, and in cases when changes need to be reverted.

[apollokit]

Haven't looked at this properly but one quick comment: these two changes are sufficiently independent to warrant 2 PRs. That's good practice and helps with debugging, and in cases when changes need to be reverted.

Yeah, I thought about that. Seems like more effort than it's worth though. The changes are well divided by the commits, and the commits are working code, so reverting changes for debugging is not a concern IMO.

Shrug. Could do it if you feel strongly about it.

[Achllle]

For a small PR it's not a big deal, but I always see people stretch the limits of what should be one PR. I think as a maintainer we should (try to) set the example.
One other advantage: it makes PRs go in MUCH faster, especially since one small change is not holding back another unrelated change that could help someone already if released.
If you don't mind ;)

[apollokit]

Per discussion, separated this pull request out into two, see the links above