A-GPS SUPL library

[pemar13]

Hi!

I use ncs v1.3.0 and mfw 1.2.0 on a nRF9160 DK v0.8.5.

My app runs in PSM mode alternating between LTE and GPS following Jan-Tores suggestions in his response to my Ticket 254606.

1. Things work reasonably well, but I am surprised by the log output from supl_session (library nrf9160_libsupl-v0.6.0-aecb5b4) when using a SIM from Telia in Sweden:

MCC location info injected, MCC: 240
	Lat: 5851520 Lon: 780606
	Unc semiminor/semimajor: 119/119
	Confidence: 100
	Orientation: 0

I assume the location is generic, based on the MCC code. However, the MCC code 240 is for Sweden whereas the location is in Hornnes in Norway!

Do you have some explanation for this?

2. Also, when using your LTE Link Monitor, the serving station location is displayed as:

MccMnc:     24001
Operator:   Telia Sverige AB
CellID:     2568dddd
TAC:        ddd

which using LocationAPI is identified as the correct station that I am connected to.

Is it somehow possible to inject this location to the GPS?

3. At some occasions, the supl_session has timed out resulting in an error. In these cases, my app just hangs until the GPS-attempt times out.

Have you ever made similar observations, and if so, is there something I can do to let the code proceed without the A-GPS functionality?

Best regards,

Per

[Heidi]

Hi!

1. Looks like you already got an answer.

2. Injecting alternative data into the GPS isn't something we have tried but it wouldn't gain anything. The position received from the SUPL server is just an alternative to the GPS having to download the same data from the satellites itself.

2. Unfortunately, we don't support the user injecting their own data into the GPS. It is something we are looking into, whether it would be beneficial to support this or not. Injecting accurate information into the GPS would indeed help shorten the TTFF.

3. Yes, I've seen this. This can happen when the SUPL session is running in the main thread resulting in the main thread not being available to read the GPS messages coming from the modem. This will fill up the memory shared between the application and the modem, so the IP packet from the SUPL server doesn't get read and the session times out.

Try running the SUPL session in a separate thread. 

[pemar13]

Thank you Heidi!

Regarding Q2: Are you saying that the time to first fix (TTFF) is totally independent of the approximate location provided during the SUPL request, which means that TTFF is identical whether or not the cellular network provides any MCC/MNC and CellID (such that Q1 is basically irrelevant from an AGPS point of view)?

Best regards,

Per

[Heidi]

Hi, I just got a reply from one of our GPS experts. My answer to your second question was incorrect, I have now updated my previous comment to reflect the correct answer, which is that we don't support the user injecting their own data into the GPS.

[pemar13]

Thanks for the clarification!

/per