Hey Matthias,

What the algorithm at that point does is check if the original trace
might contain a U-Turn at that point. If so it does not add separate
candidates for each direction (but bidirectional edge based nodes),
resulting in the possibility of U-Turns at that position.

Hope this helps.

Best,
Patrick

On Wed, Jul 22, 2015 at 11:28 AM, Matthias Schwamborn
<[email protected]> wrote:
> Hi all,
>
> looking at the code in plugins/match.hpp [1], I noticed that candidates
> resulting in a U-turn are allowed but wouldn't you say that these
> candidates are actually pretty unlikely compared to candidates that
> don't result in a U-turn? Am I missing something here? Thanks.
>
>
> Best, Matthias
>
> [1]
> https://github.com/Project-OSRM/osrm-backend/blob/master/plugins/match.hpp#L104
> --
> Matthias Schwamborn
>
> University of Osnabrück         Tel.:   +49-541-969-7167
> Institute of Computer Science   Fax:    +49-541-969-2799
> Albrechtstr. 28                 E-mail: [email protected]
> D-49076 Osnabrück, Germany      http://cs.uos.de/schwamborn/
>
>
> _______________________________________________
> OSRM-talk mailing list
> [email protected]
> https://lists.openstreetmap.org/listinfo/osrm-talk
>

_______________________________________________
OSRM-talk mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/osrm-talk

Reply via email to