Worldlight: mastering regional traffic signal control with discrete world models

Pengyong Wang; Yang Zhang; Zhiheng Li

doi:10.1117/12.3032937

5 July 2024 Worldlight: mastering regional traffic signal control with discrete world models

Pengyong Wang, Yang Zhang, Zhiheng Li

Proceedings Volume 13184, Third International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024); 131846I (2024) https://doi.org/10.1117/12.3032937
Event: 3rd International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024), 2024, Kuala Lumpur, Malaysia

Abstract

Traffic signal control (TSC) improves traffic efficiency by coordinating traffic signals across intersections. Recently, there’s a trend to apply Deep Reinforcement Learning (DRL) to TSC. However, many current DRL methods struggle to effectively coordinate intersections as they rely on training policies conditioned on both their own observations and those of neighboring intersections. This approach leads to limited information extraction, subpar performance, and an inability to adapt to changes in neighboring policies. In this paper, we propose WorldLight, a DRL method that incorporates the learning of world models. Specifically, for each intersection, we model its neighbors’ influence as its own underlying traffic flow and train a latent dynamic model with discrete representation to summarise historical observations and predict future underlying traffic flow. Additionally, we leverage Monte-Carlo Tree Search (MCTS) purely in the latent space of world models to enhance sample efficiency. Extensive experiments demonstrate the better performance of our method and better adaptability to varying neighboring policies.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Pengyong Wang, Yang Zhang, and Zhiheng Li "Worldlight: mastering regional traffic signal control with discrete world models", Proc. SPIE 13184, Third International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024), 131846I (5 July 2024); https://doi.org/10.1117/12.3032937

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
6 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Education and training

Mathematical optimization

Network architectures

Statistical modeling

Roads

Safety

Signal attenuation

Show All Keywords

Keywords/Phrases

Search In:

Publication Years