Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [sumo-user] Issue of Vehicles Stopping at Intersections
By default. vehicles will try to avoid colliding on intersections. Thus, they may stop even when their signal is green if there is still obstructing traffic on the intersection.
If you want to avoid this behavior, you can change your junction type to "traffic_light_unregulated" but be aware that this will degrade the realism of your model.

Am Di., 15. Okt. 2024 um 08:45 Uhr schrieb 최석환 via sumo-user <sumo-user@xxxxxxxxxxx>:

Dear SUMO users,

Hello, I am currently learning reinforcement learning (RL) algorithms for traffic signal optimization.

I am utilizing the Ingolstadt21 network from the RESCO benchmark (2021 NIPS) for traffic signal optimization. When I run the SUMO simulation without applying an RL algorithm, it operates normally. However, when I apply RL algorithms like MPLight or IDQN, I observe that vehicles stop at specific intersections even though signals are assigned, causing the state to continuously increase.
The green squares in the attached image represent vehicles that are stopped despite having been assigned a signal, and their vClass is bus.

  1. The lanes that the stopped vehicles are supposed to enter have enough space.
  2. The RL algorithm allocates the signals, and the simulation changes the signals properly.
  3. No special events like traffic accidents have occurred.


RESCO benchmark URL : https://github.com/Pi-Star-Lab/RESCO

I am using Python 3.9 and SUMO version 1.20.


I am curious to understand why this phenomenon happens. I appreciate your kind response in advance.

Best regards,
SeokHwan, Choi

image.png

_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user

Back to the top