Open Cornslakes opened 4 days ago
Hi there,
I've worked on a similar project and found some repositories that might be helpful for you:
ReinforcementLearning-TrafficLights: This repository provides an example of controlling traffic lights using reinforcement learning.
deeprl_signal_control: This repository also focuses on traffic signal control using deep reinforcement learning.
Deep-QLearning-Agent-for-Traffic-Signal-Control: This repository explains traffic signal control using deep Q-learning.
I hope these resources help you get started with your project. They should work well with SUMO. If you have any questions, feel free to ask!
Thanks.
Have a look at: https://arxiv.org/abs/2311.09851 DeePC cc: @AlessioRimoldi
Here to help if needed!
Hey together,
I am currently working on a university project where the goal is to develop a reinforcement learning agent which controls the traffic light states. I heard that SUMO is the most appropriate for it as the a-b-street API isn't very supportive towards versatile access to traffic lights states, agent positions etc.
Has someone already worked on some similar project and can maybe help me get a starting point on how to best develop a "light-weight" approach? It's not a big project but the topic seems interesting to me.
Thanks for any help in advance!
SUMO-version:1.20.0
operating system: WIN11