Re: [sumo-user] Demand modelling with detector traffic data given from t

[Jakob Erdmann <namdre.sumo@xxxxxxxxx>]

Sumo routes are edge based and not lane based so you cannot easily force vehicles onto specific lanes using xml input (though it is possible with departLane, arrivalLane and waypoints).

However, routeSampler can enforce turning ratios when given edgeRelation data (x vehicles going right, y vehicles going left).

You could use sumolib to analyze all the connection targets for a network lane and generate edgeRelations  for that target with to the vehicle count on that lane whenever there is only one target edge (a dedicated turning lane).

regards,

Jakob

Am Di., 17. Nov. 2020 um 15:06 Uhr schrieb Marc Schulz <schulzma@xxxxxxxxxxxxxxx>:

Dear SUMO-Team,

 

We do have detector-data for every lane on the edge. So we look after a demand file which is lane based and not edge based. The reason behind that is, that we need exactly the ratios of vehicles which drive left and which drive straight ahead. For example in our case we need to simulate that 60% of the vehicles drive to the left and 40% straight ahead.

We started by generating a demand file by using the tool “routeSampler.py”. Our input was a route-file and an edgedata-file. The route-file was generated by “randomTrips.py”. The edgedata-file was generated with the tool „edgeDataFromFlow.py“ which considered the detector-data from the traffic office (format: .csv). The output was correct if you only look for the traffic volume edge-based, but as mentioned above we need lane-based traffic flows.

 

How would you solve this kind of problem? Or are there any similar SUMO examples?

 

Thanks in advance,

 

Marc Schulz

_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user