incorrect driving orientation leading to infinite route recalculation

[sanderr]

Description

Occasionally, while driving, OsmAnd seems to get confused as to the driving orientation. The arrow starts pointing backwards or sideways. This is then followed by a route recalculation, where the navigation sends you back to the nearest roundabout/crossroad to make a U-turn, presumably because the app thinks you're going the wrong way. This then repeats every 10 seconds or so. This makes navigation next to impossible.

This seems to be tied to certain roads, but I can't say for sure at the moment.

How to reproduce?

Unfortunately I'm not in a position to investigate how to reproduce this consistently. I'll try to describe the circumstances under which I encountered this as detailed as possible.

• I've first encountered this issue about a week ago
• I've encountered it on at least six separate occasions now, four of which were on the same road (two out of two in each direction). So it seems to be tied to specific roads, but I lack sufficient data points to say for sure.
• I'm on vacation in Portugal. I've never used OsmAnd here before, and I've never encountered this issue in Belgium where I often use OsmAnd.
• I'm using Android Auto, which I've never used before, and will not have available once I return to Belgium. I include this for completeness but I don't think this is related to the issue. The first time I encountered the issue I decoupled my phone from Android Auto, recalculated the route, and got the same result.

Your Environment

OsmAnd Version: OsmAnd+ 4.1.11 Android/iOS version: Android 10 Device model: Motorola Moto G8 Plus

Maps used (online or offline):
Portugal 2022-05-01

[Gbernardy]

The same drawback also occurs with navigation with gpx track, if you deviate from the route (by mistake or on a closed road) the recalculation proposed by Osmand is never correct (#14436 and #14174).

What I do not understand is that this error has been reported for some time, for a navigator, after drawing the route from-to, the recalculation is one of the most important functions, and despite repeated reports not only is not resolved, but not even included in a milestones 🤔.

[sanderr]

I was aware of the gpx behavior, but the core of this issue is not only about how it recalculates but about the fact that it recalculates at all. I have not deviated from the route at all, yet suddenly, it decides to recalculate, possibly because it thinks I've turned around. This is probably a duplicate of #14351. For some reason I didn't find it or I dismissed it yesterday.

[vshcherb]

If you can provide a route with start - end point and we can reproduce it with Route Simulation that would be wonderful cause then we know how to reproduce and fix it. With 4.2 version we have more Route simulation options such as 'Realistic', 'Fixed speed'

[sanderr]

I didn't know about route simulation, that's interesting. I'll give that a try, but it might be a couple of days before I find the time.

[EssBee59]

Hello,

I encountered last days a similar problem: I navigated using a track, but after 2.375 km I deviated on a path (till 100 meters from the track). After nearly 50 meters on the "path-deviation" Osmand started recalculations and again and again...

Not a real problem, just the repeated calculation announcement is too fast (after each 7 -or 10 seconds) It would be nice to repeat only each 30 or 60 seconds!!!!

To reproduce the situation I will attach: -a screenshot of the location (red is the track used for navigation, blue is the path where I waIked really) -the track that I intended to follow in the Osmand-navigation -a mini "path-deviation" track (manually created with the Brouter-debug.log) -the brouter-app was used in Osmand for routing, so I provide the brouter-debug.log ==> each recalculation is protocolled, see the "from" position, example "wp=188477810,140159875,from,0.0" but this position must be converted to obtain the real coordinates) Bad_Soden_amTaunus-_10_km.gpx.txt Path_deviation.gpx.txt debug.txt

[sanderr]

If you can provide a route with start - end point and we can reproduce it with Route Simulation that would be wonderful cause then we know how to reproduce and fix it. With 4.2 version we have more Route simulation options such as 'Realistic', 'Fixed speed'

I'm afraid I couldn't manage to reproduce it with route simulation. I tried a couple of different routes to no avail.

[EssBee59]

"simulation" is possibly not exact same as life!

As explained, it is only the minor problem, that the voice-message "route recalculation..." (do not know exactly in englisch) is repeated too fast ... If you find a way to repeat only each 60 seconds, it would be fine for me! Regards

[sanderr]

If you can provide a route with start - end point and we can reproduce it with Route Simulation that would be wonderful cause then we know how to reproduce and fix it. With 4.2 version we have more Route simulation options such as 'Realistic', 'Fixed speed'

I'll provide a start and end location anyway, in case anyone else wants to give it a go: (start 37.09102, -8.41264; end 37.08893, -8.42457).

@EssBee59 I'm not sure your issue is related actually. I see what you're trying to say but I think it's only related to a side effect of this issue. This is really about the fact that it recalculates when not required, and when it does it infers the wrong driving direction.

[EssBee59]

OK, in my case the recalculation is ok, as I was outside the track! (only the frequence of the voice messages should be lower ) Abot the "wrong driving direction" I created an issue some time ago.. possibly the same problem? https://github.com/osmandapp/OsmAnd/issues/11218

regards

[sonora]

An annoying side effect is that this leads to endless repeated voice prompts "Route recalculated.. time, distance." every few seconds. This may be a possible regression, I believe that there was a time where we limited repeated route recalculated prompts to one per minute or so?

To reproduce, I drive on a 2 lane road, then deliberately miss a turn-off which OsmAnd calculates to be your best route. Once you keep straight, OsmAnd will every few seconds recalculate the route, it will be pointing behind in the opposite direction, and every time you will hear the "route recalculated, distance, time" prompt anew.

EDIT: I am not 100% sure my way to reproduce is related to the original issue (see comment below), but it could well be...

[sanderr]

While I agree that that could also be improved upon, I think it is more in line with what EssBee59 reported 4 comments up than it is with the core of this issue, where no actual deviation from the route is made at all.