Does RTKLIB have a place in the PX4 world? It’s an open-source that allow RTK levels of GPS accuracy, albeit with some possibly-taxing computing requirements. Seems as though it would be a low-cost alternative to GPS receivers with RTK built in. RTKLIB would need a certain amount of support on the groundcontrol end, but you would need two expensive GPS modules to do RTK otherwise, instead of two cheap ones. Does anyone know of attempts at running RTKLIB with PX4?
@bazeemuth Not that I am aware of, but would be something cool to have!
I did some more research (found the correct search terms!) and it turns out that RTKLIB is well-known in the Pixhawk community, dating back at least four years. In fact, Emlid’s “Reach” RTK receivers USE a fork of RTKLIB for their onboard processing: https://github.com/emlid/RTKLIB
Chris Anderson did a blog post in May 2015 about Emlid’s Indiegogo for Reach (https://diydrones.com/profiles/blogs/affordable-new-rtk-gps-designed-for-pixhawk-apm) and touted it as a lower-cost option for doing RTK positioning on a drone. And it was low cost, relative to the other GPS receivers that would do RTK at the time, and still is (except in comparison to the uBlox M8P receiver, I guess). But the Reach M+ is $265, and you need two of them if you’re not getting corrections from NTRIP (a third-party live-data source), and there are receivers priced around $90 that could provide the correct outputs to allow an external RTKLIB-based process to generate RTK positioning. To me, siting that process on Pixhawk/PX4/Nuttx makes sense if there’s enough computing horsepower there. Perhaps the transmission of RTK correction-data from the base over MAVLink would be the biggest hurdle to overcome.