From the Desks of ROS Maintainers: A Survey of Modern & Capable Mobile Robotics Algorithms in the ROS 2

Hi all, your friendly neighborhood navigator here!

I’m happy to announce yet another great paper that was recently accepted for publication! You can find the preprint at the link below (might be the last that comes from my side for awhile, but we’ll see :wink: ).

This paper reviews the algorithms, capabilities, and systems within Nav2 and the mobile robotics ecosystem in ROS 2. Its aim is to provide some expert insight from the experts themselves regarding these algorithms and advice to practitioners about what algorithms to use when (and why). It also includes formal benchmarking of performance of each of the classifications of algorithms such as planners, controllers, and path smoothers. I’m thankful for the help of @automatom, @DLu, @mikeferguson and @amerzlyakov to write up a bit about their expert opinions!

Personally, I think this will be of massive utility for folks over the next many years - moreso than the Marathon 2 demonstration of the system. It came out great with important high-level info that’s obvious to us as the maintainers and developers, but not to many of our users. Make sure to skim Appendix I for high level summaries of our recommendations. I highly recommend giving this a read if you’re using Nav2 currently or evaluating it in the future to see what capabilities you have access to and what we recommend for your application.

And of course, if you use Nav2 in your work, please cite it as:

S. Macenski, T. Moore, D. Lu, A. Merzlyakov, M. Ferguson, "From the Desks of ROS Maintainers: A Survey of Modern & Capable Mobile Robotics Algorithms in the Robot Operating System 2", Robotics and Autonomous Systems 2023.

You can find the official copy here:

Happy Nav2ing,



Nice to see mention of Fuse, does this mean fuse is now ready for general use? My understanding was it’s been WIP since 2018.

1 Like

It’s been ready for general use for a couple years now. Here’s Stephen Williams’ ROS World 2021 talk on it:

It’s also been ported to ROS 2 by the very talented folks at (formerly) OSRC, and bloomed for ROS 1 and 2 distributions.


Wow, I’m blown away that already nearly 500 people have clicked on the link to read the paper! Thanks everyone and seems we struck a chord with that manuscript!


OTTO replaced robot_localization with Fuse years back, it’s run a few million hours in production environments, plus more in test + sim.

1 Like

Its also imminently on the agenda to replace R_L starting in 2024 as the official closely integrated state estimation system - both Tom (author; maintainer) and I are on board with deprecating R_L in favor of it.

The main open question on it is GPS support continuation, the rest are just details about making an orderly transition.


This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.