This object is in archive! 

"Routing Service can't find end point. Please redefine"

Ian Kidd shared this problem 17 months ago
Solved

Hi


Please can anybody advise why I am getting the following error message

"Routing Service can't find end point. Please redefine"

Replies (11)

photo
0

Hello,

you have probably set some routing (e.g. bike) and the point is in the area where no bike path goes.

You need to change routing (manual will work always) or find the closest point but on the road.

If it doesn't help, please send me a screenshot and the coordinates of that place.

Regards,


Zdenek, Locus team

photo
1

An other workaround to get a navigation calculation is to select a different location as destination. Sometimes it helps to select a location on a street/path that is nearby.

Keep in mind that the navigation calculation is dependent on the navigation service that you are using (Google, Brouter, Graphhopper, LoRouter online/offline etc.).

Not all navigation services know the same locations and how to navigate there.

Hope this information helps.

photo
1

I have this same issue for major US highways using the "car" profile and offline LoMaps. I am not sure I can trust the router if it doesn't know about US Highway 101 going through Santa Barbara, CA. Maybe this is a bug vs. a lack of data?

You can simulate it by routing from US 101 in Santa Barbara to US 101 in Ventura. It tries to detour you into the mountains instead of following 101, and if you try to add a waypoint on that section of 101, it gives the above error.

photo
1

Just to be considered: if openstreetmap data lacks in quality for a certain location (not connected road parts, special attributes for certain vehicles, road restrictions) it is also possible to have symptoms like this (no routimg possible or strange routings that make no sense). If you have access to OSM, you could create a task for a location or try to find out whats wrong and correct it.

photo
1

Just tried it on the web version of OSM and it works fine. Definitely wondering if there is a bug/glitch.

photo
1

it may help the devs to export a GPX file of the calculated route & attach here

photo
1

I tried it out, gpx attached.If you investigate by trying small parts of routings in the problematic section (not taking the coast route), you see that probably OSM data is the cause for this particular problem and probably of the original described problem in this issue. In the screenshot you can see that there is a small gap in the highway displayed south-east of the bridge over the Carpinteria Creek direction south and if the destination point is set right behind this gap, the routing algorithm does not want to cross that gap.

So this is just the beginning of investigation about the OSM data. As the wrong routing is probably occuring at several other places on the costal route, it could be need some effords to correct it for OSM contributors. Or the LoRouting needs to be adapted to the not optimal OSM data. Graphhopper as Routingservice in Locus does not show the problem.

photo
1

Hi, this needs a deeper investigation of the OSM data behind and its use by LoRouter. For the time being, switch to GraphHopper in Locus settings > navigation > router. GH renders the route normally along the 101.

photo
1

I did some more research im OSM (by comparing the 101 at the different places) and found that in the name of "ventura highway" the 101 has a classification of parametet "vehicle" (in German screenshot: "Fahrzeuge") which looks not valid against the usual values explained in https://wiki.openstreetmap.org/wiki/Key:vehicle

I will now see if I can change the status of this parameters to a commonly accepted value ("openly accessible" or something like this should be the correct value). But as other routing services rely on other parameters to see if a street can be used, probably LoRouter developers can have a look at how to handle this example the best way (as it could ocvir for oter locations, too).

photo
1

This is well spotted druki. The primary problem is the unfortunate, unsupported and indeed totally wrong tag motor_vehicle=no|yes|yes. At the moment I am not able to tell how exactly is this "|" sign which normally means OR handled deep inside programs that build routing data. I will have a look as I want to understand this problem into detail. But it may be easier just to delete 24 usages or this mapping worldwide.

photo
1

or learn how to contribute to OSM. it's not that hard, and everything is done by volunteers. The routing quality is only as good as the underlying data allows

photo
1

on the other hand, why is online & offline routing different?

photo
1

I can not confirm a difference between LoRouter offline and LoRouter online in this case. Both take the "mountainroute". But earlier in thread it was mentioned "web version of OSM" works. So this is an other routing logic as the LoRouter.

photo
1

Both online and offline LoRouter and external Brouter have the same behavior - take the mountain road.

photo
1

Sorry, yes, the way I tested online was via https://www.openstreetmap.org/ - it routes along the coast as expected with a start of santa barbara, ca & and end of ventura, ca.


Unfortunately my main requirement is offline routing/creation/editing that works in unknown areas (not this particular area, this was just a test run.)

I think it may be that anything openstreetmaps based will be unacceptable unless I ran into the one extremely popular highway with bad data.

photo
1

@demca:
you should note that this is an outdoor app. For car it could be so it doesn't work perfectly. These bugs should always have a low prio. What is important is that cycling and hiking works well.
I always use Google Maps for car. They are at car specialists. Locus also offers a quick switch from Locus to maps. I use that often. Find me a free parking (Locus shows that) near my tour and say navigate with Maps.
The other way around also works very well now. Find me on the road a cafe on Maps and share it with Locus.

photo
1

Believe it or not.. Locus is actually one of the best apps for car (or motorcycle, in my case) offline usage. the key is OFFLINE.. much of the entire western US has poor to no cell coverage, so you cannot rely on having connectivity. Many times I plan a route in the morning, end up 300 km away and realize I need to replan my route with no signal.


Google maps is almost the exact opposite of what I want: online, fastest routing, doesn't follow the route I want it to (usually backroads/curvy roads/dirt roads.) You can download chunks of offline map data, but it is clunky and not easy to get all the chunks you need if you're crossing 1000s of kms.


Unfortunately the only real alternative are Garmin dedicated devices, and their products get worse every year. But their maps and POI are pretty solid. Ideally Garmin should just license their map data to someone else, so they could write a decent Android/IOS app!

photo
1

oh dear, always thought the USA is a very progressive country. With motorcycle knows user 0709 quite well. I think he takes kurviger.de for planning and then navigates with Locus. He is also in the forum, possibly times write to.

photo
1

@demca - this particular problem should be solved both for online and offline routing soon, just by the App update. If there is more obvious blunt errors like this, please do report them.

photo
2

I have noticed that druki has been making some edits in the OSM db, fixing this motor_vehicles=no|yes|yes nonsense. This is always a good thing to do. Anyway I reviewed how car access is treated in the car profiles and changed it in this way: If there is some unknown, unsupported value, the car access is treated by default as YES. Previously it was NO. As unsupported values are not common for roads, the problem has not yet manifested - or has not been reported. Routing using the 101 highway now works as expected for cars using the testing environment. Monday next week online routing is planned to be published, offline will follow. Thanks for reporting this tricky bug and please keep reporting obvious errors like this.

99d6abddcee2e211e5477a3f9a0fc77a

photo
2

Awesome Radim, thank you for jumping on it.. I'll keep testing it out, I would sure rather give Locus money than Garmin.. :)

Replies have been locked on this page!