TESEO: 3D-Fix, wrong position at low SNR

[Stefan Walter]

Dear all.

I am working together with a company facing a problem with TESEO-GNSS-Receiver in Japan. In an automotive usecase, with low SNR, the receiver gives us a wrong position, we've seen it off by up to 200 Meters. This alone would not be too bad, but together with the wrong position we get Fix=3D and low DOP-Values, so we must trust the position and we get lost in the countryside while actually driving on an Expressway. The GPS-Position is typically following the actual position in parallel, but with large offset. As i know this is kind of hard to understand, i'll attach a screenshot of Teseo-Suite pro with a recorded dataset. Map shows positions far east of the Expressway, SNR for all 17 Sats in use are around 30, best 38, lowest 19

To my regrets, we do not get the information, which chip exactly they use and what the configuration is like.I got some hints, though, from analysis of field data:

• we see very precise position even with low SNR (< 30dB) most of the time
• we see the problem does only with low SNR, never when SNR is > 33dB
• we see altitude is fluctuating rather strong in such cases
• we never see any Fix other than 3D or None, no Time-Fix or 2D-Fix ever
• GPS and GLONAS are used together, no other systems
• DeadReckoning is not used in the TESEO

Of course we told our customer that the SNR is very bad and leads to unreliable results, but the product is in the market so they don't want to change anything. So the only option is to distinguish between "good" low SNR 3D-Position and "bad" low SNR 3D-Fix Position - or to fix something in the TESEO-configuration. In the post https://community.st.com/t5/interface-and-connectivity-ics/teseo-liv3f-uses-3-satellites-but-does-not-get-a-position/m-p/186402/highlight/true#M2507 i read there os a note about disabling 2D-Fix. Maybe our customer has done so and thus implied "any fix > no is reported as 3D-Fix" in the firmware? Maybe there is someone else out there using a TESEO and has seen such wong 3D-Fix position before? Any hint is warmly welcome!

Thank you in advance, best Regards

        Stefan

[Stefan Walter]

... i got a few more details now. The Firmware seems to be a customer-specific version of 4.5.16 (i had to erase the Vendor-Name, sorry), and also got the configuration settings.

$PSTMVER,BINIMG_4.5.16_XXX_ARM*28
$PSTMSETPAR,1200,0x09631240*5B
$PSTMSETPAR,1227,0x02004001*52
$GPTXT,XXXXX 3.76 for v4.5.16.0*1F

Following the um2399-st-teseo-iii-binary-image--user-manual-stmicroelectronics.pdf, the config parameters look reasonable, and - as assumed - 2D-Fix is Disabled. I'd be glad is someone could explain what exactly happens if 2D-Fix is disabled and the reception situation becomes degraded. Will the fix-quality stick to 3D-Fix as long as possible, or drop down to no fix? We see "No Fix" only in parking garage/tunnel/under bridge.

In case we cannot make the Teseo confess that it has a bad position by revoking the 3D-Fix, may it be feasible to post-filter possible wrong positions based on the covariance-values in the vendor-specific $PSTMPV-Messages? I have the feeling that they become higher in the faulty conditions compared to good conditions, but would there be a reasonable threshold? Btw: is there a view in Teseo-Suite to display the covariances graphically?

All suggestions welcome.

Best Regards

        Stefan