Clarification about GNSS location hint fields

RE: Clarification about GNSS location hint fields

JONATHAN LL — Wed, 07 Dec 2022 10:18:49 GMT

We can check what is going on if you can provide us with a modem log and the coordinates that you are injecting.

https://infocenter.nordicsemi.com/topic/ug_trace_collector/UG/trace_collector/intro.html?cp=10_3_3

Regards,
Jonathan

RE: Clarification about GNSS location hint fields

jessyan — Mon, 05 Dec 2022 16:10:33 GMT

Without the location hint, we are seeing cold starts in about 30s which matches the nrf9160 spec.

With the location hint, we could be waiting for minutes before getting a fix, and most of the time never getting one after waiting 5 minutes.

RE: Clarification about GNSS location hint fields

JONATHAN LL — Fri, 25 Nov 2022 11:31:11 GMT

[quote user="jessyan"]we are actually taking longer to find a fix than if we did not have the location hint[/quote]

how much of a difference is it?

RE: Clarification about GNSS location hint fields

jessyan — Tue, 22 Nov 2022 19:45:23 GMT

Thank you for the response! I have been testing the GPS location hint with the equation you've given, using the `nrf_modem _gnss` module. However, we've noticed that with the location hint, we are actually taking longer to find a fix than if we did not have the location hint so we think the conversions could be off.

Can you confirm that the following are the correct equations and units for this location hint struct? The accuracy is in meters.

Latitude, longitude are scaled by 1e5
unc_semimajor = log(acc)/log(1.1)
unc_semiminor = log(acc)/log(1.1)
unc_altitude = 255 (we don't have altitude accuracy at the moment)
confidence = 68

Thank you!

RE: Clarification about GNSS location hint fields

JONATHAN LL — Tue, 04 Oct 2022 10:47:33 GMT

Hi,

[quote user=""]Is there a way to translate those into unc_semimajor, unc_semiminor, and confidence to pass to the GNSS module as a location hint? [/quote]

The accuracy you get form Google can be converted to K value. Accuracy is the same as uncertainty.

Just needs to be converted into unc_semimajor and unc_semiminor using the formula provided. Location is usually given with 1-sigma confidence, so should set the confidence field to 68 (i.e. 68%).

Since cellular position accuracy is given in radius of circle the unc_semimajor and unc_semiminor will be the same value.

That gives the K = Log(Accuracy)/Log(1.1), so if we have a accuracy of 300 meter the the K value is 36.

The following formula should work in Excel: =LOG(<accuracy>/10+1,1.1). depending on what version of Excel you have the "," might not be accepted as the divider for number and base.

Regards,
Jonathan