Nitpick: I don't think you need the " here and you can remove it from all CMakeLists.txt files (maybe in a separate small cleanup PR)

Not to select pose that planner gets hard to calculate path is better for perception pipeline. Because we are measuring for middleware, not for planner. Should i add this part to documentation, @sjahr ?

I think it's odd to just see the w component here. Can orientation be left empty altogether and default to the identity orientation?

Maybe you can add a small section here how to visualize the results (You can also open an issue and do that in a follow-up PR)

I actually implemented visualizer script in python. It would be great to prepare visualization relavant stuff in follow-up PR. But if you wonder it, i run benchmarks separately per each middleware (rmw_zenoh, rmw_cyclonedds and rmw_fastrtps). Then i move this files to where the script is located and run scripts. Box plot is created magically after running script. Script is [here](https://gist.github.com/CihatAltiparmak/899090aca71223fbb964569f83265a49).

Does the selected middleware have any impact on the success / failure of the planning request? I'd assume it shouldn't and in that case these numbers might be not super relevant for the benchmark. What do you think?

Thank you! IMO, this explanation should be documented in the launch file. Otherwise 👍 from me