You can find cadyts input files in
https://github.com/eclipse/sumo/tree/master/tests/tools/assign/calibration
If you put together a simple input example where flowrouter fails I can
take a look.

regards,
Jakob

Am Do., 21. März 2019 um 16:10 Uhr schrieb Kevin Malena <
[email protected]>:

> Thank you for your answer Jakob.
>
>
>
> Obviously I do know that this problem is highly underderminated and
> complex, but even for just single examplary routes flowrouter (and
> especially it’s the respect-zero function) doesn’t really do what I want it
> to.
>
> In combination with the calibrators it works pretty well, but I thought
> there could be a more elegant solution which I don’t see because I am very
> new to SUMO.
>
>
>
> Maybe u or someone else can explain me how to create a <REALFLOWS>-file in
> cadytsIterate.py because i wanna give it a try just to compare the results.
>
> The <REALFLOWS>-file has to be a XML-file in form of:
>
> <measurements>
>
>     <singlelink link="1to21" start="25200" end="32400" value="750"
> stddev="8" type="COUNT_VEH"/>
>
>     <singlelink link="1to22" start="25200" end="32400" value="250"
> stddev="8" type="COUNT_VEH"/>
>
> </measurements>
>
>
>
>
>
> I converted my detector data to the following format which is compatible
> for flowrouter:
>
>
>
> Detector;Time;qPKW;vPKW
>
> Detektor_Nummer_16_unten;0;2;10
>
> Detektor_Nummer_1_unten;1;1;11
>
> Detektor_Nummer_11_unten;2;2;13
>
> Detektor_Nummer_2_unten;0;3;9
>
> Detektor_Nummer_8_unten;0;2;10
>
> Detektor_Nummer_13_oben;1;1;12
>
> Detektor_Nummer_7_oben;1;2;12
>
> …
>
>
>
> With help of the additional file I can also figure out the lane and exact
> position of those defined detectors.
>
>
>
> I suppose the <REALFLOWS>-file in cadytsIterate.py uses aggregated data
> for the time span start to end on the different lanes, but I am not sure.
>
>
>
> Thanks & Cheers,
>
>
>
> Kevin
>
>
>
>
>
>
>
> *Von:* [email protected] [mailto:[email protected]]
> *Im Auftrag von *Jakob Erdmann
> *Gesendet:* Donnerstag, 21. März 2019 13:44
> *An:* Sumo project User discussions
> *Betreff:* Re: [sumo-user] Validation of detected data
>
>
>
> The problem flowrouter has to solve in a city network is strongly
> underdetermined. There are plenty of knobs and dials to improve the result
> (among other things you can forbid explicit routes O/D-pairs.
>
> You could also try cadyts which adapts demand levels to detectors for a
> pre-determined set of routes:
>
> https://people.kth.se/~gunnarfl/cadyts.html
>
> https://sumo.dlr.de/wiki/Contributed/Cadyts
>
>
> https://www.researchgate.net/publication/318418803_SUMO-Cadyts_calibration_with_limited_data_quality
>
>
>
> regards,
>
> Jakob
>
>
>
>
>
> Am Mi., 20. März 2019 um 13:21 Uhr schrieb Kevin Malena <
> [email protected]>:
>
> Hey…
>
>
>
> I want to validate detected data in a sumo network of my city (at
> different spots i placed sensors to detect every incoming and outgoing
> vehicle with speed, arrival time and type).
>
> I triedto use flowrouter, but the generated routes don’t really fit the
> data.
>
> Right now I implement a combination of the generated routes and flows by
> flowrouter with the ‚dynamic‘ calibrators using route probes (see docu:
> Building a scenario without knowledge of routes, based on flow
> measurements).
>
> Do you know about a different method like flowrouter which is more precise?
>
>
>
> Thanks,
>
> Kevin
>
> _______________________________________________
> sumo-user mailing list
> [email protected]
> To change your delivery options, retrieve your password, or unsubscribe
> from this list, visit
> https://www.eclipse.org/mailman/listinfo/sumo-user
>
> _______________________________________________
> sumo-user mailing list
> [email protected]
> To change your delivery options, retrieve your password, or unsubscribe
> from this list, visit
> https://www.eclipse.org/mailman/listinfo/sumo-user
>
_______________________________________________
sumo-user mailing list
[email protected]
To change your delivery options, retrieve your password, or unsubscribe from 
this list, visit
https://www.eclipse.org/mailman/listinfo/sumo-user

Reply via email to