Hello, - using e3 detectors this way is fine. - the 'freq' attribute changes the aggregation frequency. - meanTimeLoss * vehicleSumo is mostly correct. It also includes slow-downs due to turning angle which you would even get if the light were green all the time. - yes. Only vehicles that entered and left are counted. - the *within* values will be included again in later intervales when these vehicles actually leave the detection area so you can ignore them if you plan to look at all intervals.
regards, Jakob Am Mo., 10. Mai 2021 um 09:57 Uhr schrieb Menno van der Woude < [email protected]>: > Hello all, > > suppose I have a signalized intersection, and wish to know the delay > caused by the traffic light, what would be a good approach? > > Currently I do this: > > - place e3-detectors with entries before the intersection, and exits > after the intersection > - side note: this leads to warnings in SUMO cause I place > entries/exits of traffic that is diverging from or converging to a > single > lane on that lane; SUMO shows a lot of warnings, telling me about > vehicles > that only entered or left the e3 detector. is it bad practice to place > e3 > detectors this way? > - collect data (can the 15 min default interval be changed?) > - then use meanTimeLoss * vehicleSum as the total delay that occured > for a given route over the intersection during the given interval > > Is this a viable approach? Will it correctly estimate the delay caused by > the traffic lights? > > A few specific questions: > > - am I right to assume that only vehicles that entered *and* left the > e3 detector will be counted in "vehicleSum"? (and time loss...) > - should I take into account "meanTimeLossWithin"? or will > "meanTimeLoss" for the next interval reflect the "time loss within" from > the previous round? currently I am assuming the latter. > > Thanks for any thoughts and tips! > > Greets, > > Menno van der Woude > CodingConnected e.U. (Wien - Österreich) > E [email protected] > W www.codingconnected.eu > _______________________________________________ > sumo-user mailing list > [email protected] > To unsubscribe from this list, visit > https://www.eclipse.org/mailman/listinfo/sumo-user >
_______________________________________________ sumo-user mailing list [email protected] To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user
