Add map gaps FAQ item #180
Comments
|
Here they are all the same tower, just like it is swimming and it needs a few gasps of air every few meters. A mystery. Samsung A13 5G. |
|
I didn't observe issues like this on any of my phones. The only way to diagnose is to review the logs from the app gathered at detailed level. If you share some by email I will take a look. |
|
I'm going on a long trip (which is the perfect time to gather such stretches) at the end of the month. So what do I do? Do I set this to debug? And then CSV files get bigger and I send you one of those? Or maybe the reports you want are sent automatically by email according to this setting?: |
|
You'll need to send the log file by email. I suggest to compress it first which should significantly reduce the file size. The second option only reports then application crashes and doesn't include the execution details. |
|
OK I am analyzing the logs! Here is the picture: |
|
Long (A), and short (B, C) gap repeating pattern, https://en.m.wikipedia.org/wiki/Yucca,_Arizona . The gaps seem to be totally unrelated to where cell ID changes. Just like the phases of the moon are unrelated to when Venus will rise. Here after about three A, B, C's the cell ID might change, and not in any particular spot. |
|
Without logs on debug level it's not possible to say where the issue comes from. I'm using Tower Collector on 5 different phones and I didn't observe such behavior, so my conclusion is that's device specific problem. |
|
Yes, the higher the speed of the bus/train , the clearer the patterns. (Except I don't recall it happening with 2G GSM originating from within the same airplane.) Anyway yes, perhaps Samsung A13 5G specific. It would be best for someone who makes daily trips and sees the same |
|
If this is related to speed then or may be either poor GPS signal quality resulting in low accuracy or cell towers switch with a delay (at least phone delivered interrupted data). |
|
This time the train was going 90 mph instead of 79, so the gaps got stretched out so we can see them in more detail. That part of Arizona was a perfectly flat desert with ideal cell towers. The gaps are not related to cell tower handoff or anything like that. GPS was perfect 2 m resolution, in a glass roofed Amtrak observation car. Anyway indeed I did submit logs a long time ago, back in #180 (comment) , please have a look. (b.txt.gz attachment.) |
|
I missed that attachment, thanks! It looks like there was no update from GPS in given time but it was reconnected just before that period. It seems you have And one question - do gaps appear all the time or nearly to places when train speed changes significantly? |
|
OK, I turned it off and next trip I'll see if it will help. The gaps show up best when the train is traveling at a high constant speed in a straight line. |
|
"GPS optimizations" seems to be the critical item. Turning them off seems to fix the problem for me. Therefore there should be a help item for it. It should list the pros and cons. Saying why one should pick it, and why one might not want to pick it. As far as the pros, well I'm not sure what they are. Maybe it is for saving electricity. Be sure to mention them. As far as the cons go, mention that from around 50 to 100 miles per hour, gaps appear on the map every few miles, on e.g., some Samsung phones. As far as why gaps do not appear at commercial airline speeds, I suppose that the plane is going so fast that distance is no longer a factor, and then the one second time check becomes the factor. And each time the time check occurs the distance is already beyond the minimum needed and a point is recorded. As far as high speed rail bullet trains speeds are concerned, I didn't test them. @zamojski |
|
@zamojski Hope you got my latest comment. |
…on some devices when moving at high speeds. For zamojski#180. (Which @zamojski might be missing notifications of). Yes, tracing through the code, one finds ``` transportMode = (MyApplication.getPreferencesProvider().getGpsOptimizationsEnabled() ? MeansOfTransport.Universal : MeansOfTransport.Fixed); ``` and in CollectorService.java, things like ``` Timber.d("updateDynamicInterval(): Skipping GPS reconnection because of too small interval difference: %s", intervalDiff); ... int newInterval = Math.min(currentIntervalValue.get() + 500, transportMode.getMaxTime()); ```
|
Yes, that PR is to help people who see these gaps,
|
|
The solution should be to validate if it still makes sense to use GPS optimization on modern devices and why it causes such issues. When I will be working on a new version I will take a closer look at it. |
|
Indeed, looking at https://github.com/barbeau/gpstest , I see |
Here is a typical railroad trip (US. railroad max. speed 79 mph.) There are these odd gaps, almost regularly, and not exactly at tower handoffs either. The handoffs often occur a few dots later, but sometimes don't occur at all.
Perhaps you can analyze your own train and road trips, and author an FAQ item, saying what's going on. Thanks.
And no, it's probably not related to GPS satellites. Because only once in many months, did I notice a GPS glitch on the map like this, very rare:
The text was updated successfully, but these errors were encountered: