RobustLight: Improving Robustness via Diffusion Reinforcement Learning for Traffic Signal Control

Mingyuan Li, Jiahao Wang, Guangsheng Yu, Xu Wang, Qianrun Chen, Wei Ni, Lixiang Li, Haipeng Peng

TL;DR: Improving Robustness via Diffusion Reinforcement Learning for Traffic Signal Control

Abstract: Reinforcement Learning (RL) optimizes Traffic Signal Control (TSC) to reduce congestion and emissions, but real-world TSC systems face challenges like adversarial attacks and missing data, leading to incorrect signal decisions and increased congestion. Existing methods, limited to offline data predictions, address only one issue and fail to meet TSC's dynamic, real-time needs. We propose RobustLight, a novel framework with an enhanced, plug-and-play diffusion model to improve TSC robustness against noise, missing data, and complex patterns by restoring attacked data. RobustLight integrates two algorithms to recover original data states without altering existing TSC platforms. Using a dynamic state infilling algorithm, it trains the diffusion model online. Experiments on real-world datasets show RobustLight improves recovery performance by up to 50.43\% compared to baseline scenarios. It effectively counters diverse adversarial attacks and missing data. The relevant datasets and code are available at Github.

Lay Summary: Traffic lights controlled by AI reduce congestion, but attacks or missing data can disrupt them. We propose RobustLight, which is a smart fixer that fixes corrupted data in real time. It recovers the corrupted data and improve the performance of traffic signal control, keeping traffic flowing without replacing existing systems.

Application-Driven Machine Learning: This submission is on Application-Driven Machine Learning.

Link To Code: https://github.com/AdvancedAI-ComplexSystem/SmartCity/tree/main/robustlight

Primary Area: Reinforcement Learning

Keywords: reinforcement learning, diffusion, traffic signal control

Submission Number: 1225