Questionable time conditions of induction loop detector (E1)

[m-kro]

Discovered thanks to a request from the mailing list:

• The function traci.inductionloop.getLastIntervalVehicleNumber will not always return the vehicle count from the last interval, but sometimes less
• The enter/exit times are computed differently compared to the "instant induction loop" (delta = 1 TS) and to what the visualisation in sumo-gui suggests

Can be reproduced using the example from the mailing list and

• using detector period of 1 TS
• printing the corresponding TraCI functions traci.inductionloop.getIntervalVehicleNumber & traci.inductionloop.getLastIntervalVehicleNumber for every time step

Essentially substracting 1 TS at some points would correct this, but I doubt a bit given the error persists since several years at such a crucial point...

[m-kro]

Let us wait for @behrisch to return

[YuccaTang]

I am wondering any progress about the bug. The request link above is wrong. I attached the correct one below.

https://www.eclipse.org/lists/sumo-user/msg13258.html

Based on my investigation, the bug in Traci has nothing to do with time conditions. It may be induced by vehicles suddenly changing to other lanes when they are just on E1 detectors.

[YuccaTang]

@m-kro I hope my finding may help you.