Problems with LoRouter and estimated arrival time

[Vinterblad]

I have Locus Maps 4 Gold. Whenever I try to use LoRouter to navigate it fails miserably to give me a realistic estimate of time of arrival. As far as I can see its almost always double the time. There is specifically two long distance drives I do 4-8 times a month and I know the routes by heart. I still use navigation for the time estimates so I can give realistic information on time of arrival to workmates and friend. This problem forces me to use other GPS software and I would like to avoid that.
The shorter route is, according to LoRouter 669 km and its estimated time of travel is 12:28 (normally takes just over 6 hour). The longer route is, again according to LoRouter 1210 km and the estimated time to arrival is 22:35 (takes 11 to 13 hours depending on food stops).

Anything I'm doing wrong with settings or something that is easy to fix?

[janaton]

Hello Vinterblad,
please, could you share one of the planned routes with us (feel free to ignore some distance around start and destination to keep your privacy)? It is hard to investigate without further details. Thank you very much.

Kind regards,
Ondrej

[Vinterblad]

Hello and thanks

Is the gpx-file, exported from the navigation in Locus, enough? I also added a screenshot that shows the estimated travel time and the distance for the attached gpx-file.

Best regards

[Graf Geo]

I can confirm some nonsensical times of estimated arrival on navigation.

Example: trip from Strandbad Wannsee (Wannseebadweg  25, Berlin) to Coubertinplatz, Westend, Berlin.
Trip goes through Grunewald forest in Berlin.

With LoRouter offline (profil names in german):
Gehen (shortest): 13,8 km, 1:14 hs
Gehen (LoWalking): 15,3 km, 1:22 hs
Wandern: 15,7 km, 1:24 hs
Radreise: 16,6 km, 1:15 hs
Rad Straße: 16,4 km, 1:15 hs
MTB: 13,9 km, 1:03 hs

With BRouter offline:
Gehen (shortest): 13,8 km, 2:45 hs
Wandern: 16,6 km, 0:58:41 hs
Radreise: 16,5 km,  1:05 hs
Rad Straße: 16,6 km - 0:58:41 hs

You see, all walking/hiking times are bullshit , only BRouter with profil "shortest" calculates realistic time.

[abrensch]

I can also confirm:

LoRouter's car-routings (both fast and eco) seem to use a fixed speed of 50 km/h for ETA

It's not using BRouter's ETAs from the kinematical model

[Menion]

Thanks to changes made by @Radim V and later detected also in the official BRouter repo, times will be updated in the next Locus Map 4 (Beta) version.

[Graf Geo]

I don't have the beta version, but the changelog for LM 4.1.0 states, among other things: "More precise ETA values in the LoRouter". So I checked this yesterday and today.

For me, a reasonable ETA is important when walking and hiking. Often we want to reach a rarely running train or bus at the end of the tour, and there it is important to know if you have to hurry or not.

I usually use track navigation with a route created with the route planner. With V 4.0 this worked well last time, the ETA always predicted the arrival time quite accurately.

Now with LM 4.1.0 (and 4.1.1) this no longer works.

A track with e.g. 10 km is estimated with 1h:46m at the beginning. (walking)
A track with e.g. 5 km is estimated with 57m:48s at the beginning. (walking)
This corresponds to a speed of approx. 5.65 km/h or approx. 5.15 km/h. That is clearly too fast for walking (I am not in the army). You can try this out on all kinds of tracks, it is never correct and the time is always calculated far too close!

Even worse: There is also no proper adjustment to the actual speed during the tour.

Yesterday I let myself navigate along a test route of 12.4 km created with the route planner. At the beginning, Locus estimated 2h:15m! (corresponds to approx. 5.51 km/h).

I then took a leisurely walk, a few short breaks, met a familiar person on the way and chatted for 10 minutes. The route was recorded for comparison:

Track time 3h:04m, in motion 2h:44. (so about 20 minutes breaks).
Speed (total): 3.9 km/h
Speed (in motion): 4.4 km/h

Every 1km I took screenshots and logged the displayed values (ETA and remaining time). The ETA continuously shifted backwards, but only reached the realistic time very shortly before the finish.

Excerpts from the logging:

Start 17:16, remaining distance 12.4 km, time to finish 2:15, arrival time 19:31.
(Here you can see immediately that you can't do this!)

Check 18:37, remaining distance 7.2 km, time to finish 1:22, arrival time 19:59
(Here you can see immediately that you can't do this!)

Check 19:27, remaining distance 4,1 km, time to finish 0:47, arrival time 20:13

Actual arrival 20:20.
For the full log, see the attached excel sheet.

Locus calculates with an assumed speed of approx. 5.2 - 5.6 km/h over the entire distance and does not take into account that the actual speed is significantly slower!

In my opinion, the ETA calculation does not work properly at the moment and is therefore unusable. Then I can and must calculate a realistic arrival time in my head based on the remaining distance.

What do you mean?

*** Translated with www.DeepL.com/Translator (free version) ***

[Graf Geo]

Update: hmm, may be that the router is slowly "learning".

In the routes created with the route planner, the time calculated during creation is somehow saved. I created the same route the day before yesterday exactly the same way again. This time the time is calculated with 4.7 km/h. (For the first route it is 5.7 km).
When saving the route, these values are obviously saved.

I have now saved the absolutely identical route of 10 km twice, once created the day before yesterday (A) and once created a few minutes ago (B).

If I want to navigate it, 1:46 hours are estimated for the older version A and 2:08 hours for the new version B. Still too little, but already much better.

Let's see if the calculation improves further in the next few days....

[Graf Geo]

After several tours, the ETA is now better, but it is absolutely intransparent and not comprehensible what is calculated and how. Routes created with Webpaner result in a different, often nonsensical ETA than routes created with LM4.
A description or documentation would be helpful here.

[Graf Geo]

And now only nonsense is being calculated again. For the same distance of 5 km, which this morning was realistically calculated 1h:04m, suddenly 32m:46s are estimated again.
The ETA calculation is totally screwed up.

[Menion]

Thanks for the discussion guys and Graf Geo for the topic on the help desk.
If it is not a problem, let's continue with it there: https://help.locusmap.eu/topic/24639

[Graf Geo]

Dear Locus team,

may I ask you to have a look at the helpdesk thread https://help.locusmap.eu/topic/24639, especially my last post from today.

There I show a very simple example of how a single shaping point causes a completely wrong ETA calculation. I hope this helps to understand the problem.

Thank you very much!

[Menion]

Work in progress. I'm marking it as "Solved" here and we may continue in the Help Desk if there are any more issues.

3 years to solve this? Not bad 

[Graf Geo]

... 3 years to solve this? Not bad 

I thought exactly the same thing. 

Thanks.