Re: [sumo-user] Deadlocks in junctions

[Daniel K. <d.k...@go...>]

Dear Niclas,

well, I talked to Jakob and we think that:

The behaviour displayed in picture 1 is very surprising. I would assume
that you use a non-default car-following model. Is that true? I suppose
this should not occure with Krauß.

For the second picture, the intersections must be joined. I tried to
implement something that solves such situations, but dropped it due to its
complexity. A vehicle would have to look forward whether it is blocked
and/or blocks other vehicles. This is not straightforwards, as the vehicle
would have to extrapolate his leaders actions into future as they may brake
due to some reason forcing him to stop, too. This hardly works. So we
decided to not cover such cases - they do not resemble real world
intersections anyway. So you have to join them.

We do not yet know why joining does not work in your case, we'll try to
find out in the next time. You can try to add a nodes fiels to your
netconversion configuration where you join intersections manually, like:
<join nodes="42735222 42735014"/>

Of course putting your node ids into the nodes attribute.

sincerely,
Daniel


2012/3/22 Niclas Raabe <nic...@ud...>

> Here are the links to the pictures:
>
> Picture1:
>
> http://postimage.org/image/fx3x5fy8z/
>
> Picture2:
>
> http://postimage.org/image/rpk9tfz2t/
>
> Greetz
> Niclas
>
> Am 21.03.2012 um 13:35 schrieb Niclas Raabe:
>
> > Hello!
> >
> > I'm using SUMO to simulate the traffic of the Ruhrgebiet in Germany. So
> I took the OSM-Data of an extract of the road network and converted it via
> NETCONVERT to a SUMO network file with the following command:
> >
> > ./netconvert --osm-files ruhrgebiet.osm -o ruhrgebiet.net.xml
> --junctions.join-dist 30 --ramps.guess --tls.join --remove-edges.isolated
> --no-turnarounds
> >
> > If I raise or decrement the distance of --junctions-join there is no big
> difference between the results. Even if I raise the distance to 390 there
> are junctions to see in Picture1.png. If you have a look at this picture
> you can see two lanes with waiting cars. In this simulation this cars are
> waiting the whole simulation. Can anybody tell me why they are waiting and
> not driving? In my opinion there is no reason not to drive into the
> junction for both cars that are the first in the lane.
> >
> > In Picture2.png you can see a similar situation of the same simulation.
> The route of the left car in the junction is colored in ochre. Why is this
> car not driving over this junction? As far as I can see there is no car in
> the way. Or am I wrong?
> >
> > I don't use teleporting at the moment. In my opinion there has to be a
> way not to use it to solve this situations.
> >
> > I hope you can explain me this behavior of the cars.
> >
> > Greetz from Germany
> > Niclas
> >
> >
> >
> > <Picture1.png><Picture2.png>
>
>
>
> ------------------------------------------------------------------------------
> This SF email is sponsosred by:
> Try Windows Azure free for 90 days Click Here
> http://p.sf.net/sfu/sfd2d-msazure
> _______________________________________________
> sumo-user mailing list
> sum...@li...
> https://lists.sourceforge.net/lists/listinfo/sumo-user
>