Hi Gervalino,
it obviously depends where the pedestrians walk in your network. You can
observe the pedestrian in question using sumo-gui to check where the angle
changes drastically. It could be some intersection where they turn.
Currently pedestrians can make quite abrupt changes and travel along rather
simple trajectories when using the internal movement model for pedestrians.
Do you run sumo with very short time steps? What would you expect to be an
acceptable change in orientation for pedestrians?
Best regards
Mirko
-----Original-Nachricht-----
Betreff: [sumo-user] Pedestrian Orientation - Moments when there are some
"jump" in this information.
Datum: 2023-11-22T19:55:54+0100
Von: "Gervalino Bilas via sumo-user" <sumo-user@eclipse.org>
An: "Sumo project User discussions" <sumo-user@eclipse.org>
Hello.
I'm running the code below to simulate 20 pedestrians walking through an
area to get position (x,y) and orientation of them every single second.
lstPeopleId = traci.person.getIDList()
for personId in lstPeopleId:
x, y = traci.person.getPosition(personId);
orientation = traci.person.getAngle(personId);
The code is running well and I can export a CSV file to work with that
data.
But something about the orientation data is making me worried. There are
some moments when there are some "jumps" in this information.
For example, in a second the user's orientation is 251.56° and in the next
it is 340.73°.
The same behavior it's not found in location data - it changes less
abruptly.
So my question is: There is a formula (logic or algorithm) that rules how
orientation information changes?
My goal is to understand how exactly this information is generated. If
there are a scientific article explain that will be perfect.
PS: Check an file with 50 regs attached where is possible to see "the
jumps"
Regards,
_______________________________________________
sumo-user mailing list
sumo-user@eclipse.org
To unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/sumo-user
