An adaptive traffic light control system using reinforcement learning
Traffic signal control (TSC) is a challenging issue in managing an urban transportation system. A fixed time TSC is easy to implement but has drawbacks in such measures as flow rate, waiting time, and traffic density. The situation gets worse when the arrival rates of vehicles periodically change...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Prince of Songkla University
2022-08-01
|
Series: | Songklanakarin Journal of Science and Technology (SJST) |
Subjects: | |
Online Access: | https://sjst.psu.ac.th/journal/44-4/2.pdf |
_version_ | 1797848263313850368 |
---|---|
author | Kietikul Jearanaitanakij Chanayut Jamkhaw Nattapat Puangpipat Tot Worasrivisal |
author_facet | Kietikul Jearanaitanakij Chanayut Jamkhaw Nattapat Puangpipat Tot Worasrivisal |
author_sort | Kietikul Jearanaitanakij |
collection | DOAJ |
description | Traffic signal control (TSC) is a challenging issue in managing an urban transportation system. A fixed time TSC is
easy to implement but has drawbacks in such measures as flow rate, waiting time, and traffic density. The situation gets worse
when the arrival rates of vehicles periodically change over time, which is usual in most urban cities. We propose adaptive
reinforcement learning (RL) to manage TSC with varying vehicle arrival rates. Our objectives are to improve the averages of
flow rate and waiting time and reduce the wasteful green light problem by considering the vehicle densities of the current lane
and the downstream directions. Experiments were conducted by Simulation of Urban MObility (SUMO) under three traffic
layouts and various vehicle arrival rates. The proposed method not only reduced on average traffic density, waiting time, and
queue length, but also increased the average flow rate and average speed, relative to the other algorithms tested. |
first_indexed | 2024-04-09T18:24:39Z |
format | Article |
id | doaj.art-b86005252d8b4137a0369d22b9694a6b |
institution | Directory Open Access Journal |
issn | 0125-3395 |
language | English |
last_indexed | 2024-04-09T18:24:39Z |
publishDate | 2022-08-01 |
publisher | Prince of Songkla University |
record_format | Article |
series | Songklanakarin Journal of Science and Technology (SJST) |
spelling | doaj.art-b86005252d8b4137a0369d22b9694a6b2023-04-12T04:22:30ZengPrince of Songkla UniversitySongklanakarin Journal of Science and Technology (SJST)0125-33952022-08-0144491492210.14456/sjst-psu.2022.122An adaptive traffic light control system using reinforcement learningKietikul Jearanaitanakij0Chanayut Jamkhaw1Nattapat Puangpipat2Tot Worasrivisal3Department of Computer Engineering, School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Lat Krabang, Bangkok, 10520 ThailandDepartment of Computer Engineering, School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Lat Krabang, Bangkok, 10520 ThailandDepartment of Computer Engineering, School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Lat Krabang, Bangkok, 10520 ThailandDepartment of Computer Engineering, School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Lat Krabang, Bangkok, 10520 ThailandTraffic signal control (TSC) is a challenging issue in managing an urban transportation system. A fixed time TSC is easy to implement but has drawbacks in such measures as flow rate, waiting time, and traffic density. The situation gets worse when the arrival rates of vehicles periodically change over time, which is usual in most urban cities. We propose adaptive reinforcement learning (RL) to manage TSC with varying vehicle arrival rates. Our objectives are to improve the averages of flow rate and waiting time and reduce the wasteful green light problem by considering the vehicle densities of the current lane and the downstream directions. Experiments were conducted by Simulation of Urban MObility (SUMO) under three traffic layouts and various vehicle arrival rates. The proposed method not only reduced on average traffic density, waiting time, and queue length, but also increased the average flow rate and average speed, relative to the other algorithms tested.https://sjst.psu.ac.th/journal/44-4/2.pdftraffic signal controltransportationreinforcement learningadaptive green light timewasteful green light problem |
spellingShingle | Kietikul Jearanaitanakij Chanayut Jamkhaw Nattapat Puangpipat Tot Worasrivisal An adaptive traffic light control system using reinforcement learning Songklanakarin Journal of Science and Technology (SJST) traffic signal control transportation reinforcement learning adaptive green light time wasteful green light problem |
title | An adaptive traffic light control system using reinforcement learning |
title_full | An adaptive traffic light control system using reinforcement learning |
title_fullStr | An adaptive traffic light control system using reinforcement learning |
title_full_unstemmed | An adaptive traffic light control system using reinforcement learning |
title_short | An adaptive traffic light control system using reinforcement learning |
title_sort | adaptive traffic light control system using reinforcement learning |
topic | traffic signal control transportation reinforcement learning adaptive green light time wasteful green light problem |
url | https://sjst.psu.ac.th/journal/44-4/2.pdf |
work_keys_str_mv | AT kietikuljearanaitanakij anadaptivetrafficlightcontrolsystemusingreinforcementlearning AT chanayutjamkhaw anadaptivetrafficlightcontrolsystemusingreinforcementlearning AT nattapatpuangpipat anadaptivetrafficlightcontrolsystemusingreinforcementlearning AT totworasrivisal anadaptivetrafficlightcontrolsystemusingreinforcementlearning AT kietikuljearanaitanakij adaptivetrafficlightcontrolsystemusingreinforcementlearning AT chanayutjamkhaw adaptivetrafficlightcontrolsystemusingreinforcementlearning AT nattapatpuangpipat adaptivetrafficlightcontrolsystemusingreinforcementlearning AT totworasrivisal adaptivetrafficlightcontrolsystemusingreinforcementlearning |