So, you are 4.7 m (15’5”) tall? Now we know that Belgians are no short people…
But seriously, somehow I have been expecting better, especially from Galileo. It is like 30 years younger than GPS. I was hoping for finally less than 1 m ground accuracy for civilian use, at least that is what we have been promised by EU politicians and ESA officials. However, it is interesting that your device catches satellites just 6° to 9° over the horizon. This is no easy feat!
Anyway, in your tests, did you disable Wi-Fi, Bluetooth, and cell phone network positioning?
I also recently replaced an 8 year old single-band phone (Galaxy S7) with a year old dual-band phone (Pixel 7). Its current-gen GNSS chipset should be comparable to Samsung’s, although it’s Broadcom rather than Qualcomm (not that I think anyone could tell any difference in consumer-grade electronics). I’m in NE Europe, so satellite coverage is also similar. I’ve only been on a couple trips so far, but the overall resulting GPX track seems to be noticeably better for me than before. I imagine it’s a combination of newer GNSS chipset, software processing and of course dual-band. This is somewhat anecdotal evidence because I haven’t done any side-by-side comparison tests, but I did go through the recoded tracks and compared their accuracy/precision and noise to my older tracks. One thing that immediately stands out is that it’s much better in obstructed and partly-obstructed environments, like passing between tall buildings, forested areas, under bridges, etc. My old GPS would jump all over the place and sometimes just fly off to Narnia. The new GPS remains almost as accurate as unobstructed view. This does match the “advertised” benefit of dual-band being able to filter out GPS reflections/errors. And I think there’s some corrective/predictive software processing involved (may be even using compass) as it sometimes slightly overshoots, but it is also able to follow small changes really well. I saw almost no difference in accuracy/precision in perfect conditions with unobstructed views (at least nothing I can tell without side-by-side comparison). However, it is immediately noticeable how any obstruction would cause old GPS to deviate while the new one remains accurate. Basically, under ideal conditions when GPS isn’t having any issues, it’s roughly the same as it was before. But it’s having a lot fewer issues now, which was the main problem for Mapillary recordings since most of the “important” locations are not obstruction-free.
As an example, here’s a location along a tall building with old and new GPS:
Sometimes having everything you can get is not the best option. In my experience, usually Wi‑Fi and Bluetooth positioning introduces more noise into GPS data than any improved accuracy. Yet, it may be helpful as a fallback option. However, 5G positioning is supposed to be quite good from what I hear and read because it has been intended to work especially in buildings and have shopping isle in supermarkets like accuracy. I have not tried it yet, due to lack of a 5G and network coverage.
One thing that immediately stands out is that it’s much better in obstructed and partly-obstructed environments, like passing between tall buildings, forested areas, under bridges, etc.
This does match the “advertised” benefit of dual-band being able to filter out GPS reflections/errors.
It is great to hear that dual‑band works as advertised in obstructed environments. Urban canyons are definitely an issue for Mapillarians.