L76K filtering out inaccurate acquisitions

IoT Modules GNSS Module
1

Dear Support,

The product I am working on contains L76K GPS module. The observation is that GPS module ends up in one of three possible “states”:
(1). GPS signal quality is good, fix is obtained quickly, reported location is accurate
(2). GPS signal quality is bad, fix is not obtained at all, no location is reported
(3). GPS is in obstructed environment, so the fix is obtained, but reported location is inaccurate

The logic we would like to apply is to filter-out (3) to behave like (2), because our product logic prefers to not have location at all rather than have inaccurate location.

Unfortunately, L76K does not seem to expose any direct information about the GPS fix quality, like accuracy (for example exposed via PQEPE command described in “Lx6&LC86L&LG77L Series GNSS Protocol Specification” (unfortunately not present on L76K))

We have tried using Dillution-of-precision (“HDOP” field in GGA sentence) or number-of-sattelites (“NumSatUsed” field in GGA sentence), however they seem to be unreliable for detecting (3), i.e. we observed (3) happening with low HDOP and high number of sattelites.

As a part of L76K cold and hot start for low-power periodic location acquisition - #2 by nkastelik I have received reference manuals for L76K, but haven’t found any information there regarding this particular case.

Questions:

  1. Is there any recommendation on how to detect NMEA sentences that report GPS fix, but with low accuracy (i.e. case (3)), so they can be rejected ?
  2. Is there any recommendation on how to improve GPS accuracy in general, so that potentially (3) could become (1) ?
  3. Is there any possibility to configure GPS quality threshold in L76K, so that (3) are reported as (2) by the L76K module?

Best regards,
Nikodem

2

Hi @nkastelik

I am sorrry for the late reply. The answer about your question as follows:

Q1: Is there any recommendation on how to detect NMEA sentences that report GPS fix, but with low accuracy (i.e. case (3)), so they can be rejected ?
A1: As you said, the accuracy of position is high relative with signal quality. So, I recommend you through your software aspect to calculate the avaerage CN0 of all satellites, which every satellites’s CN0 is indicated by the parameter like of GSV. Then, evaluate the average CN0 to judge the position of this time whether trust, like trust the position when average CN0>=35.

Q2:Is there any recommendation on how to improve GPS accuracy in general, so that potentially (3) could become (1) ?
A2:This module is old, so its function already fixed. So, i recommend use external software program to make up shortcoming.

Q3: Is there any possibility to configure GPS quality threshold in L76K, so that (3) are reported as (2) by the L76K module?
A3: I am sorry to say no.

Best regards