GPS based parking mode: camera randomly wakes up

[tchavei]

Hi.

So I decided to test the A139 Pro for a few days before installing it into my car.

Today I connected it to a usbc charger 5V, 3A and just place it on my balcony, on the sun, to see how much it would warm up etc.

I left it for about 3h. After 5 min, the camera correctly enters parking mode (no GPS movement). I've set parking mode to AED, movement sensitivity low, g sensor sensitivity to high.

When I returned, I had 16 parking events because my cats like to visit the balcony far more often then I expected. It was a good test. Camera detected and recorded the cats perfectly. It didn't lock any file as none of my cats bumped into it (damn. Maybe next time).

However, I noticed that the camera 'woke up' randomly at least 3 times with `one channel recoding started' message and then recorded continuously during 5 minutes until it went back into parking mode.

I wonder why it happens. Maybe a GPS drift?

If it's really a momentarily GPS drift that makes the camera think it's moving, I'm not worried as I will use the HK3C kit in the car but if it isn't gps drift, then I'm worried...

Anyone had a similar experience or has any idea what might be happening?

Thank you

 

[tchavei]

Well, I guess I'm answering myself on this one too...

Yes. It was GPS drift. Kind of weird though. I analysed one of the clips where the camera wakes up from parking mode and the previous clip (20 min earlier or so). There's a sudden jump in coordinates and the camera registers an "apparent" speed of 8km/h for about 10 seconds. That obviously cancels parking mode.

Since I'm going to use the HK3-C, I'm safe I think.

However, it seems to me @viofo could change the firmware a bit to ignore these small random drifts. The coordinate jump was momentarily but the recorded speed was constant for 10 seconds. I don't know how viofo does the speed pooling but it seems it needs some adjustment.

I bet this would help immensely all those users out there that only can have GPS based parking mode for whatever reasons.

Cheers.

 

[Dex_]

Cheap GPS chipset cause it. Viofo long time use GAM-2525 MTK3337 for A129, A129Plus, A229 dashcam products. This chipset works with GPS&QZSS satellites only (no Galileo, no Glonass, no BeiDou) without any features like DGPS or FW update. Default settings can't be changed. Update data frequency no more than 1Hz.
I do not have A139Pro so I don't know that chipset used for it but I'm pretty sure that it is still same.
Then car moving slowly on parking zone on many video records shown 0 Km/h. And then car stop at traffic lights speed decrease on video with some delay ~2 seconds. So dashcam firmware already use some average values. Increase filtering makes this delay more badly.

Hope that in new models Viofo finally moves to new chipset with more types GNSS support that increase accuracy for speed and position.

 

[Classic]

GPS A139 Pro

 

[Dex_]

Thank's for photo. Same chipset as before. Outdated. 2.32 FW for this chipset was released 2015/08/31.

 

[kamkar]

I do think the GPS is probably where corners are cut / money being made.

Personally i would like to see Galileo support, CUZ why should i rely on American / Russian / Chinese when EU have their own system.

 

[tchavei]

I do think the GPS is probably where corners are cut / money being made.

Personally i would like to see Galileo support, CUZ why should i rely on American / Russian / Chinese when EU have their own system.

... And with a high precision. 10x

What does worry me is: I've been pondering if I should display the speed or not in my footage. I sometimes have a heavy foot and it wouldn't be very smart displaying the actual speed. On the other hand, if the supplied GPS isn't accurate, then a non speeding situation can become a speeding situation by error margin (which is even worse lol). Maybe I should just leave the GPS coordinates.

Of course if forensics really want to know the speed, they can calculate it from the footage but maybe displaying random speed is pretty worse, I don't know.

 

[kamkar]

Well i looked into that ( precession ) and it seem the EU system are not that much more accurate than the US GPS are,
A + with Galileo is that the way it operate, detecting jammers ASO would be faster to do.

GPS use to have a build in inaccuracy for civilian use, but that was turned off in 2000, but before that as a civilian you would have a 200 M inaccuracy

GPS: " On April 20, 2021, the global average URE across all satellites was ≤0.643 m (2.1 ft.), 95% of the time."
Galileo: < 20 CM horizontal - < 40 CM vertical. 99% of the time, this is both service level 1 & 2

Mind you Galileo full service will not be up and running before next year ( 2024 )

In regard to GLONASS and BEIDU,,,,, well i have not bothered to look into it.

 

[Dex_]

GPS, Glonass, Galileo and BeiDou have full globe coverage. As much GNSS types supported as receiver can choose more precise information from closer to zenith point satellites with high strength signal. Different satellites constellation use different frequencies bands for signal so it allow to calculate diff and fix errors like atmosphere retention and building reflections. Support many GNSS is a winning strategy and use only one of them is a bad idea.