Re: [sumo-user] fcd output and time step accuracy

[Jakob Erdmann <namdre.sumo@xxxxxxxxx>]

A main source of differences between simulations with different step lengths is vehicle insertion (more frequent insertion attempts lead to higher insertion flow). See https://sumo.dlr.de/docs/Simulation/VehicleInsertion.html

To achieve vehicle behavior closer to the 1-second simulation, run the simulation with --step-length 0.1 --action-step-length 1

Make sure that all vehicles depart at full seconds or this will be another source for scenario divergence.

Am Sa., 11. Juli 2020 um 00:47 Uhr schrieb Tetris <schmelter@xxxxxxxxxxxxxxxxxxx>:

behrisch wrote
> Am 06.07.20 um 09:24 schrieb Tetris:
>> I used fcd output to get the acceleration and speed data of a simulation.
>> The total travel time is 168 seconds longer with --step-length 1 than
>> with
>> --step-length 0.1.
>
> This sounds like a lot. What is the absolute value of the total travel
> time?

Hi,

I compared another simulation with step length 1 and step length 0.1:

<http://sumo-user-mailing-list.90755.n8.nabble.com/file/t300/steplength_comparison.png>

The vehicle in the simulation with step length 0.1 can drive while in the
step length 1 simulation it has to stand still. For example around 300
samples in the step length 1 simulation compared to arond 3000 samples in
the step length 0.1 simulation.

So the behaviour is different.

https://sumo.dlr.de/docs/Simulation/Safety.html#action_step_length
<https://sumo.dlr.de/docs/Simulation/Safety.html#action_step_length> 
The wiki says the reaction time is set by step time. What I actually wanted
was the same simulation with the same behaviour but more frequent output
values.




--
Sent from: http://sumo-user-mailing-list.90755.n8.nabble.com/
_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user