[SUPPORT] Android & iPhone app snap-to bug.

[bgodette]

Two repeatable examples, in both videos GPS Essentials was reporting < 5m for accuracy on GPS position and AGPS turned off.

Client video #1:

Trace #1: Permalink

trace1.jpg (167.93 KiB) Viewed 300 times

Client video #2:

Trace #2: Permalink

trace2.jpg (139.33 KiB) Viewed 300 times

In the second video where the position is snagging/jumping/sliding backwards early on is a third location, but there's no road off to the north or south for it to snap to so is behaving just like being "off road".

This problem does NOT exist in the Android 2.4.* clients and has been present in every Android 3.x release so far. It's unconfirmed to also be a problem on the iPhone client, but it's likely to be the same there as this would be a core routine common to both.

[troyv]

Nice experimentation!

The thing I find odd is that the green line is almost too straight. It's almost as if the point A and point B that are being used for the green line are far away from each other, maybe even up to a mile. Could it be possible that during that mile or so, that the GPS was then losing accuracy? And for some reason, the server only recorded points A and B?

Just a thought.

[bgodette]

New video with 3.1.99.121 and test road for the bug to snap to.

trace3.jpg (197.57 KiB) Viewed 250 times

Permalink
Next map update will include the test road to the south.

Again, this problem does not exist in the 2.4.x client.

[troyv]

New video with 3.1.99.121 and test road for the bug to snap to.

Again, this problem does not exist in the 2.4.x client.

I was just by there yesterday. If I get a chance to test it out, I might do that and I'll let you know what I find. I only have the released version of the app, but if it is a 3.x bug, I should be able to see it too.

What tool do you use for the capturing the screen for the video?

[troyv]

One other thing I noticed. Have you turned on the GPS Points layer? There seems to be a "black hole" of GPS points in that area.

[troyv]

Sorry for these multiple posts, but this is rather interesting. Here is another "black hole"

[jasonh300]

There are GPS black holes all over the place. I don't know if it's areas lacking gps reception or if it's a lack of data coverage that causes it.


I'm not drunk...I'm using Tapatalk.

[bgodette]

There are GPS black holes all over the place. I don't know if it's areas lacking gps reception or if it's a lack of data coverage that causes it.

The potential for lack of data coverage is why I thought this might be an issue with A-GPS, but since I have that disabled and that it doesn't do this with the 2.4.x clients sorta strikes that.

What's interesting is the two gaps in the GPS tracks on Hampden pretty much exactly correspond with where the 3.x clients go nuts.

[crazycomputingdotnet]

Makes me wonder if there is anything on or near those areas that are transmitting in the same frequency that GPS uses that is drowning out the signal. Maybe something is leaking signal in that area like a cable amp or something.

[bgodette]

Makes me wonder if there is anything on or near those areas that are transmitting in the same frequency that GPS uses that is drowning out the signal. Maybe something is leaking signal in that area like a cable amp or something.

This also occured to me, but if that were the case it would happen to the 2.4.x client as well. In addition, GPS Essentials, GNav, and Signal Finder all track perfectly through there. Also the speedometer does not have an issue which is based on measuring the doppler shift of the GPS signals' carrier frequencies, any interference would affect that as well.