Deep reinforcement learning based application in traffic signal control

• Main Author: Guo, Yi
• Other Authors: Wang Dan Wei
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Electrical and electronic engineering::Control and instrumentation::Control engineering
• Online Access: https://hdl.handle.net/10356/149618
• Institution: Nanyang Technological University

The rapid economic development has continuously improved the transportation network around the world. But at the same time, the substantial increase in vehicles has made traffic jams and traffic accidents increasingly serious. It is important to find a Traffic Signal Control (TSC) method which can be used in Intelligent Transportation System (ITS). An effective method is to use Rein forcement Learning (RL) in TSC. In this dissertation, one of the useful and easy algorithm in Reinforcement Learning, Deep Q-Network (DQN), is used to control the traffic signals. A transportation network in Singapore is built on the PTV Vissim platform and the DQN Algorithm is implemented through MATLAB. MATLAB calls the COM of PTV Vissim and conducts co-simulation with PTV Vissim. Five groups of comparative experiments are conducted with the DQN Algorithm, which has well demonstrated the effectiveness of the DQN Algorithm in reducing traffic congestion and time delay.