Citation

Du Y, Shangguan W, Chai L. Transp. Saf. Environ. 2022; 4(4): tdac027.

Copyright

(Copyright © 2022, Oxford University Press)

DOI

10.1093/tse/tdac027

PMID

unavailable

Abstract

Reinforcement learning-based traffic signal control systems (RLTSC) can enhance dynamic adaptability, save vehicle travelling time and promote intersection capacity. However, the existing RLTSC methods do not consider the driver's response time requirement, so the systems often face efficiency limitations and implementation difficulties. We propose the advance decision-making reinforcement learning traffic signal control (AD-RLTSC) algorithm to improve traffic efficiency while ensuring safety in mixed traffic environment. First, the relationship between the intersection perception range and the signal control period is established and the trust region state (TRS) is proposed. Then, the scalable state matrix is dynamically adjusted to decide the future signal light status. The decision will be displayed to the human-driven vehicles (HDVs) through the bi-countdown timer mechanism and sent to the nearby connected automated vehicles (CAVs) using the wireless network rather than be executed immediately. HDVs and CAVs optimize the driving speed based on the remaining green (or red) time. Besides, the Double Dueling Deep Q-learning Network algorithm is used for reinforcement learning training; a standardized reward is proposed to enhance the performance of intersection control and prioritized experience replay is adopted to improve sample utilization. The experimental results on vehicle micro-behaviour and traffic macro-efficiency showed that the proposed AD-RLTSC algorithm can simultaneously improve both traffic efficiency and traffic flow stability.

Language: en