Dual frequency GNSS accuracy screenshots

[Jon8RFC]

With the December 2019 update for the Pixel 4 and Pixel 4 XL, users now have access to the L5 (US) and E5a (EU) satellite signals. Let's see what you're getting for accuracy.

It seems, to me, that Android has imposed a hard-coded minimum limit on the horizontal accuracy. I initially blamed an app, but it seems like it's Android which has gone from 6ft accuracy to 9.9ft (2012-2013, if I recall correctly) to now 12.4ft (2017-2018) horizontal accuracy. That doesn't make sense to me, unless it's really just v2 of Selective Availability, now client-side. I remember reading, somewhere, about another dual frequency gnss phone receiving l5/e5a signals, but the article stated that those weren't actually being used in measurements. GPSTest shows "U" for them, so I'm not sure why accuracy is no better than before, unless the software is limiting it.

Here's the gold-standard everyone has used for a few years (it's "GPSTest", not to be confused with the countless "GPS Test" apps):
https://play.google.com/store/apps/details?id=com.android.gpstest

It's not consistent in receiving L5/E5a, and that may be because it's heavily overcast in my area of Texas today. My screenshots are from indoors, on this overcast day. "Force full GNSS measurements" is enabled in developer options, although I've not noticed any subjective difference it makes, through all of the months I've used it.
Two screenshots:
https://imgur.com/a/6W6PeKC

 

[slyyke]

Do you mind uploading your gps.conf file? And what carrier do you have?

With the December 2019 update for the Pixel 4 and Pixel 4 XL, users now have access to the L5 (US) and E5a (EU) satellite signals. Let's see what you're getting for accuracy.

It seems, to me, that Android has imposed a hard-coded minimum limit on the horizontal accuracy. I initially blamed an app, but it seems like it's Android which has gone from 6ft accuracy to 9.9ft (2012-2013, if I recall correctly) to now 12.4ft (2017-2018) horizontal accuracy. That doesn't make sense to me, unless it's really just v2 of Selective Availability, now client-side. I remember reading, somewhere, about another dual frequency gnss phone receiving l5/e5a signals, but the article stated that those weren't actually being used in measurements. GPSTest shows "U" for them, so I'm not sure why accuracy is no better than before, unless the software is limiting it.

Here's the gold-standard everyone has used for a few years (it's "GPSTest", not to be confused with the countless "GPS Test" apps):
https://play.google.com/store/apps/details?id=com.android.gpstest

It's not consistent in receiving L5/E5a, and that may be because it's heavily overcast in my area of Texas today. My screenshots are from indoors, on this overcast day. "Force full GNSS measurements" is enabled in developer options, although I've not noticed any subjective difference it makes, through all of the months I've used it.
Two screenshots:
https://imgur.com/a/6W6PeKC

 

[Jon8RFC]

Do you mind uploading your gps.conf file? And what carrier do you have?

Attached with appended .txt extension so that it would upload without having to be zipped.
Google store edition Pixel 4 on Visible (Verizon Wireless's in-house MVNO).

 

[lattosimbaste]

The GPS accuracy here post December update seems to have degraded as well. I had hoped that dual frequency performance would improve both the reliability of reception and overall accuracy but that has not been the case, while it is currently acceptable the position "walks" quite a bit so lane-level positioning seems as distant as ever.

Kalman filters seem to reduce this error but I haven't had the time to thoroughly test. I have always suspected that the inclusion of additional sensor input(barometer, gyroscope and accelerometer) for a IMU approach would assist.

Let's hope additional updates will address this including use of the Galileo GNSS constellation.