Open AccessDueling Double Deep Q-Network for Adaptive Traffic Signal Control with Low Exhaust Emissions in A Single Intersection
Author(s) -
Shu Fang,
Chen Feng,
Hongchao Liu
Publication year - 2019
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/612/5/052039
Subject(s) - intersection (aeronautics) , computer science , signal (programming language) , control (management) , traffic signal , reinforcement learning , automotive engineering , simulation , real time computing , artificial intelligence , engineering , transport engineering , programming language
In order to reduce traffic exhaust emissions caused by the large quantities of vehicles, this paper studied the traffic signal control (TSC) model with low exhaust emissions on the basis of the deep reinforcement learning. In this study, the Dueling Double DQN with prioritized replay (DDDQN-PR) algorithm we proposed was combined with the Double DQN, Dueling DQN, and prioritized replay to achieve the goal of low exhaust emissions of TSC. The agent was trained in traffic simulator USTCMTS2.1 in a single intersection. The experimental results show that the performance of DDDQN-PR was significantly better than the other four algorithms, not only in data efficiency but also in final performance.