zsoerenm
commented
10 months ago That's very cool to hear. I've never tried this on a Raspberry Pi 4 (only on a Nvidia Jetson Xavier). I'm glad that it works in real time there, too. I started out a restructure of the Tracking module earlier this year: https://github.com/JuliaGNSS/Tracking.jl/pull/43 Although it is already merged I haven't released it yet. This will make the GNSSReceiver even more flexible, because you can implement your own Tracking algorithm very easily (using multiple dispatch). I'm planning to continue this work early next year. A documentation would be very cool. I started out the Tracking documentation here: https://juliagnss.github.io/Tracking.jl/stable/ However, this isn't adapted to the merged PR that I've mentioned above.
I think JuliaGNSS is a very useful package, and it have a promise of becoming the next generation of GNSS signal processing toolbox after the Kai Borre scripts in Matlab. Its running faster than real time on Raspberry Pi 4 while the Matlab receiver runs 10x slower than real time on my desktop Threadripper PC, while its much more flexible as the GNSS receiver components are separated into its own packages.
What I think is missing is a top level documentation on how all the subpackages work together, something like https://docs.sciml.ai/DiffEqDocs/stable/ (and its source code, https://github.com/SciML/DiffEqDocs.jl), a bunch of getting started tutorials and hello worlds, like GNSS-SDR My first position fix: https://gnss-sdr.org/my-first-fix/ , and issue tracker for whole organization.
If I have time I'll probably start writing this