This object is in archive! 

Navigation - Recalculation fails

Jürgen Gärtner shared this problem 5 years ago
Solved

I went on a bicycle ride circle. After about half of the way I tried to navigate home. BRouter suggests the way back, that's ok.

Then, always after every 130 m distance the report "Recalculated" pops up. But even the preferred configuration methode is "Endpoint priority" the calculated route wrongly points backwards.

On the next mainroad I stopped the navigation and started it new. Everthing is fine now.

Maybe the configuration "Endpoint priority" is not regarded correctly.

Locus 3.31.3, Android 7.1.1

Replies (19)

photo
1

Hi,

how did you start the navigation back home? Did you use the previous route leading from home and made its reverse copy for backwards navigation?

photo
1

Hi Michal,

no, I was on the trip without active Locus and thought by myself: Let's test the navigation. So I startet

Locus | Navigation | Bicycle | Navigation To | Own Points | home

> "reverse copy for backwards navigation?" - No idea how to do this.

Jürgen

photo
1

Hi Jürgen,

I see, you just navigated home which was in your point of view "backwards". I understood your question incorrectly. As for the "recalculation" problem - did you follow the route or not? If not, that would normally trigger the recalculation. If the app recalculated despite you went along the navigated route, it was some kind of error I can't simulate now. You yourself wrote that after second start of navigation all was OK. Have you encountered the same issue after that again?

photo
1

Hi Michal,

let's have a look to the screenshots again. In the middle of this circle I started a navigation. The best way at this moment was to go back. That was correct. But I didn't followed the calculated route, instead I moved forward in reverse direction of the calculated route. Then always after 100 m Locus started a new recalculation, but not to the final destination. Instead it wants to go always back even if an obvious main road was the shortes path to destination. You see the screenshots at this point where I stopped this abnormal navigation and started a new one manually.

photo
1

Hi Jürgen,

is it possible to repeat the issue? If yes, we can do some more research what is the reason. If not, I'm sorry.

photo
1

Hi Michal,

sure, I can, Tell me what you need. Should I repeat it by bike or by car?

photo
1

It doesn't matter, any navigation profile. Any simulation of the issue is important. If we can't simulate an issue, we can't solve it.

photo
1

Hi Michal,

I went on another circle to produce the recalculation issue again, see attached screenshot.

On the first picture I've started the navigation. The calculated route was correct computed at this time.

As in the previous example I did not follow the suggested route but instead I drove forward in the opposite direction of the calculated route. Recalculation started.

On the next pictures you see the same issue as in the example before. The recalculation leeds always back but not on the shortest way to destination.

What can I do more? I think this example gives you the same information as the first example. How can I do a simulation as you requested?

photo
1

Hi Jürgen,

I think I finally understand the problem. It seems like you have strict route following turned on. Go to Locus settings > Navigation > Advanced > Strict route following > turn OFF.

photo
1

This configuration was always off and I've nerver changed it. On the other side I cannot imagine that this configuration item could be something to do with our problem.

photo
1

A shortcut or a return without halting that previously planned unidirectional (forward only) navigation track ? Autorecalculate correctly suggests, or to the next nearest forward trackpoint, OR to the next nearest unconsumed forward Via point (Finish) ! If necessary autorecalculate shall repeatedly suggest a correction (return) toward the still unfinished track section. (= The "out of bar" functionality, when user forgot where he came from). If deciding for such a very serious navigation change, FINISH and STOP the actual navigation job by the preplanned undirectional track completely. Than replan. Start up a NEW navigation to (any new POINT) position. Does not work ?

photo
1

Ah, OK, then we're at the beginning again. "Strict route following" forces user to get back to the ordinary route , resp. to the last missed routepoint. Your screenshots look like this.

photo
1

Hi Michal,

Good news!

I tried the recalculation test again. And this time all worked well.The recalculated routes were correct, as intended, as expected.

So, and now you maybe want to know what is the difference between these two cases?

On the screenshots of my second test you see on the middle picture that I drove onto the main road but the third picture shows that the recalculation never pointed to the final destination, even not to the path I came from. Instead it calculated a new route to the starting point of my navigation.

That means the error must be found at the beginning of the navigation process.

The difference is as follows:

1. Expected case, correct recalculated route:

Start Locus | wait for satellite fix | start navigation | begin movement

2. Erroneous case, incorrect recalculation:

Start Locus | start navigation | begin movement

If in the second case the navigation was started before the first satellite fix occurred, then on the upper part of the screen a small window pops up with "Wait ..". Now the localization service is in hunt-mode and the route calculation will not start before GPS fix. After the fix was happen, the starting point for calculation is given and the route calculation can be done. But now all recalculation trials will lead always back to this first GPS fix position and not to the final destination. (Maybe that is what is called "Strict route following".)

photo
1

Hi Jürgen,

thanks for reporting. So you started navigation without GPS fix - that's the moment that was missing in our testing. We'll have a look at it.

photo
1

Hi Michal,

setting the state for this PR to green "Answered" is not a good idea.

I reported a problem and I delivered the root cause too. Now it's your turn to fix it.

Otherwise we have a known bug in the system and nobody cares for it.

And don't think the reported malfunction is unlikely. Especially if you sit in a car the satellite signal is weak. Additional many people (like me too) are offline if they don't have an internet flatrate. So also the A-GPS data could not be loaded. As result the first satellite fix can be delayed up to 5..10 minutes.

All who start a navigation in this time wrongly run into the strict route recalculation method. But nearly all want to endpoint preferred.

I plan to replace the car navi (TomTom, Garmin) completely by Locus. But I want trust it seriously.

And we all want help to make Locus a great system!

Set this PR to "In Progress" and forward it to Menion. I'm sure he needs not more as 30 minutes to fix it.

photo
1

Hi Jurgen, marking a question "answered" means question is answered, not problem solved. I'll discuss it with Menion, don't worry but the usecase is quite specific so iMO it won't have too high priority. What's more, nobody else has succeded in simulating this issue yet.

photo
1

Because everybody whom was presented a stupid route thinks: "Ok, it's only a toy .."

There are not so much people who spend their time to find out what the reason is and report an error.

And, if you really want to stimulate this error, you must simply do what I have written above.

photo
1

Hi Jürgen,

I discussed this with Menion and he said he was trying to solve a similar issue two months ago, but all the same, without success of even simulating the issue. We'll be trying and testing more and also waiting if anybody else will experience this. The worst problems are those that can't be repeated on our own devices...

photo
1

Good day Jürgen,

your description of problem simulation (GPS enabled at the start or not) was not exact, but it finally helped me to find most probably the reason of this problem.

Not sure if you use Beta versions. If not, I just published version 3.31.3.9 RC into Beta channel, so you may give it a try.

Menion

photo
1

Hi Menion,

wow, everything works fine. Thank you for your work. Ok, I've tested only one situation, but this was the second example above. Now the recalculated route is the expected one. Well done!

And btw., I did not assumed "GPS enabled or not", my assumtion was "localization fix already happend or not".

You can merge it to main :-)

Jürgen

Replies have been locked on this page!