You should be able to count all maneuvers with a <change> event. Sublane maneuvers (repositioning within a lane) are intentionally excluded from the output. This may cause changeEnded events to become lost when the motivation changes during the maneuver (which is probably a bug).
Thanks for the example. Am Mi., 19. Juni 2019 um 16:41 Uhr schrieb Nico Weber <[email protected]>: > Hi Jakob, > > > > thanks for your explanation so far! Attached you find the zipped > sample-scenario (change ending from “.piz” to “.zip” to open it). Based on > your answer I try to explain my problem more detailed again: > > > > Primary, my aim is to analyse how different attributes within the traffic > simulation are distributed in case of stochastic traffic. Therefore I built > up a more complex simulation which shall approximate a stochastic traffic, > which could occur in reality. In the second attachment you find the > lanechange-output of the corresponding simulation. To analyse only the > “correct” and “successful” lane changes, for now I thought, that if the > “changestarted” and “changeEnded” options are activated, only lane > changes from vehicles in the output-file where these three change-states > occur for an specific lane change are the successful ones and therefore are > the only lane changes I can investigate afterwards further. > > > > Then I saw in the sample scenario, that a seeming “successful” lane > change can occur in the corresponding lanechange-output with only the > “changeStarted” and “change” event. Due to this fact I am unsure under > which condition I have to count a lanechange as “successful” in the more > complex simulation, because, like mentioned above, sometimes all three > states occur, sometimes not. > > > > In addition, maybe it can give you a hint that there is never the reason > “sublane” for lane changing in both, the sample and the complex simulation. > Maybe this could be caused by the fact that I chose a high value for the > lateral resolution, because I want the sublane-Model to be activated for > its additional options but I do not want to have real sublanes within the > simulation. Can you give an answer how to achieve this in general? I have > in mind the following explanation of the doc: > > > > Lane-changing model for sublane-simulation > <https://sumo.dlr.de/wiki/Simulation/SublaneModel> (used by default when > setting option *--lateral-resolution <FLOAT> > <https://sumo.dlr.de/wiki/Basics/Notation#Referenced_Data_Types>*). This > model can only be used with the sublane-extension. > > *Caution:* > This model may technically be used without activating sublane-simulation > but this usage has not been fully tested and may not work as expected. > > > > > > Thanks for your effort again. > > > > Best regards > > Nico > > > > *Von:* [email protected] [mailto:[email protected]] > *Im Auftrag von *Jakob Erdmann > *Gesendet:* Mittwoch, 19. Juni 2019 14:05 > *An:* Sumo project User discussions <[email protected]> > *Betreff:* [EXTERNAL] Re: [sumo-user] Definition/Condition > "changeStarted"/"changeEnded" event in Lanechange Output > > > > I've updated the description at > https://sumo.dlr.de/wiki/Simulation/Output/Lanechange#Continuous_Lane_Changes > > The cause of the observed behavior may be that a lane change is started > for non-sublane reasons and then the reason is changed to 'sublane' before > finishing the maneuver. This could indicate a bug. Please provide a minimal > sample scenario. > > > > Am Mi., 19. Juni 2019 um 09:44 Uhr schrieb Nico Weber <[email protected] > >: > > Dear all, > > > > i am currently writing my Master thesis regarding the safety validation of > highly automated driving functions through simulative approaches. > Therefore, amongst others, I use SUMO as tool for the traffic simulation. > In detail, I actually investigate the lane change behavior. Within the Sumo > user doc (https://sumo.dlr.de/wiki/Simulation/Output/Lanechange) the > definition of the “change” element of the lane change output is explained, > but not those of the “changestarted” events and “changeEnded” events in > case of activating this option. Could you please explain what conditions > have to be fulfilled to trigger these events? The background is, that in my > more complex simulation I saw that all three types of change events occur. > But within a simple dummy simulation, where only some lane changes are > performed, only the “changestarted” events and “change” events occur, > although vehicles perform “complete” lane changes. If necessary it is no > problem to send you the corresponding files of the dummy simulation. > > > > Thanks in advance. > > > > Best regards > > Nico > > _______________________________________________ > 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
