Deep Reinforcement Learning for Autonomous Traffic Light Control

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In urban areas, the efficiency of traffic flows largely depends on signal operation and expansion of the existing signal infrastructure is not feasible due to spatial, economic and environmental constraints. In this paper, we address the problem of congestion around the road intersections. We developed our traffic simulator to optimally simulate various traffic scenarios, closely related to real-world traffic situations. We contend that adaptive real-time traffic optimization is the key to improving existing infrastructure's effectiveness by enabling the traffic control system to learn, adapt and evolve according to the environment it is exposed to. We put forward a vision-based, deep reinforcement learning approach based on a policy gradient algorithm to configure traffic light control policies. The algorithm is fed real-time traffic information and aims to optimize the flows of vehicles travelling through road intersections. Our preliminary test results demonstrate that, as compared to the traffic light control methodologies based on previously proposed models, configuration of traffic light policies through this novel method is extremely beneficial.
Original languageEnglish
Title of host publication2018 3rd IEEE International Conference on Intelligent Transportation Engineering (ICITE)
PublisherIEEE
Pages214-218
Number of pages5
ISBN (Electronic)978-1-5386-7831-2
ISBN (Print)978-1-5386-7832-9
DOIs
Publication statusPublished - 15 Oct 2018
Event3rd IEEE International Conference on Intelligent Transportation Engineering, ICITE 2018 - Singapore, Singapore
Duration: 3 Sep 20185 Sep 2018

Conference

Conference3rd IEEE International Conference on Intelligent Transportation Engineering, ICITE 2018
CountrySingapore
CitySingapore
Period3/09/185/09/18

Fingerprint

Reinforcement learning
Reinforcement Learning
reinforcement
Telecommunication traffic
learning
Traffic
traffic
Traffic control
Infrastructure
Simulators
Intersection
Real-time
Control systems
infrastructure
road
Economics
Preliminary Test
Traffic Control
Gradient Algorithm
Urban Areas

Keywords

  • 3d Virtual Reality Simulator
  • Autonomous Traffic Control
  • Component
  • Deep Reinforcement Learning
  • Machine Learning

Cite this

Garg, D., Chli, M., & Vogiatzis, G. (2018). Deep Reinforcement Learning for Autonomous Traffic Light Control. In 2018 3rd IEEE International Conference on Intelligent Transportation Engineering (ICITE) (pp. 214-218). [8492537] IEEE. https://doi.org/10.1109/ICITE.2018.8492537
Garg, Deepeka ; Chli, Maria ; Vogiatzis, George. / Deep Reinforcement Learning for Autonomous Traffic Light Control. 2018 3rd IEEE International Conference on Intelligent Transportation Engineering (ICITE). IEEE, 2018. pp. 214-218
@inproceedings{9d817c59054e4ada85a6b469bc9933c1,
title = "Deep Reinforcement Learning for Autonomous Traffic Light Control",
abstract = "In urban areas, the efficiency of traffic flows largely depends on signal operation and expansion of the existing signal infrastructure is not feasible due to spatial, economic and environmental constraints. In this paper, we address the problem of congestion around the road intersections. We developed our traffic simulator to optimally simulate various traffic scenarios, closely related to real-world traffic situations. We contend that adaptive real-time traffic optimization is the key to improving existing infrastructure's effectiveness by enabling the traffic control system to learn, adapt and evolve according to the environment it is exposed to. We put forward a vision-based, deep reinforcement learning approach based on a policy gradient algorithm to configure traffic light control policies. The algorithm is fed real-time traffic information and aims to optimize the flows of vehicles travelling through road intersections. Our preliminary test results demonstrate that, as compared to the traffic light control methodologies based on previously proposed models, configuration of traffic light policies through this novel method is extremely beneficial.",
keywords = "3d Virtual Reality Simulator, Autonomous Traffic Control, Component, Deep Reinforcement Learning, Machine Learning",
author = "Deepeka Garg and Maria Chli and George Vogiatzis",
year = "2018",
month = "10",
day = "15",
doi = "10.1109/ICITE.2018.8492537",
language = "English",
isbn = "978-1-5386-7832-9",
pages = "214--218",
booktitle = "2018 3rd IEEE International Conference on Intelligent Transportation Engineering (ICITE)",
publisher = "IEEE",
address = "United States",

}

Garg, D, Chli, M & Vogiatzis, G 2018, Deep Reinforcement Learning for Autonomous Traffic Light Control. in 2018 3rd IEEE International Conference on Intelligent Transportation Engineering (ICITE)., 8492537, IEEE, pp. 214-218, 3rd IEEE International Conference on Intelligent Transportation Engineering, ICITE 2018, Singapore, Singapore, 3/09/18. https://doi.org/10.1109/ICITE.2018.8492537

Deep Reinforcement Learning for Autonomous Traffic Light Control. / Garg, Deepeka; Chli, Maria; Vogiatzis, George.

2018 3rd IEEE International Conference on Intelligent Transportation Engineering (ICITE). IEEE, 2018. p. 214-218 8492537.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Deep Reinforcement Learning for Autonomous Traffic Light Control

AU - Garg, Deepeka

AU - Chli, Maria

AU - Vogiatzis, George

PY - 2018/10/15

Y1 - 2018/10/15

N2 - In urban areas, the efficiency of traffic flows largely depends on signal operation and expansion of the existing signal infrastructure is not feasible due to spatial, economic and environmental constraints. In this paper, we address the problem of congestion around the road intersections. We developed our traffic simulator to optimally simulate various traffic scenarios, closely related to real-world traffic situations. We contend that adaptive real-time traffic optimization is the key to improving existing infrastructure's effectiveness by enabling the traffic control system to learn, adapt and evolve according to the environment it is exposed to. We put forward a vision-based, deep reinforcement learning approach based on a policy gradient algorithm to configure traffic light control policies. The algorithm is fed real-time traffic information and aims to optimize the flows of vehicles travelling through road intersections. Our preliminary test results demonstrate that, as compared to the traffic light control methodologies based on previously proposed models, configuration of traffic light policies through this novel method is extremely beneficial.

AB - In urban areas, the efficiency of traffic flows largely depends on signal operation and expansion of the existing signal infrastructure is not feasible due to spatial, economic and environmental constraints. In this paper, we address the problem of congestion around the road intersections. We developed our traffic simulator to optimally simulate various traffic scenarios, closely related to real-world traffic situations. We contend that adaptive real-time traffic optimization is the key to improving existing infrastructure's effectiveness by enabling the traffic control system to learn, adapt and evolve according to the environment it is exposed to. We put forward a vision-based, deep reinforcement learning approach based on a policy gradient algorithm to configure traffic light control policies. The algorithm is fed real-time traffic information and aims to optimize the flows of vehicles travelling through road intersections. Our preliminary test results demonstrate that, as compared to the traffic light control methodologies based on previously proposed models, configuration of traffic light policies through this novel method is extremely beneficial.

KW - 3d Virtual Reality Simulator

KW - Autonomous Traffic Control

KW - Component

KW - Deep Reinforcement Learning

KW - Machine Learning

UR - http://www.scopus.com/inward/record.url?scp=85057220868&partnerID=8YFLogxK

U2 - 10.1109/ICITE.2018.8492537

DO - 10.1109/ICITE.2018.8492537

M3 - Conference contribution

SN - 978-1-5386-7832-9

SP - 214

EP - 218

BT - 2018 3rd IEEE International Conference on Intelligent Transportation Engineering (ICITE)

PB - IEEE

ER -

Garg D, Chli M, Vogiatzis G. Deep Reinforcement Learning for Autonomous Traffic Light Control. In 2018 3rd IEEE International Conference on Intelligent Transportation Engineering (ICITE). IEEE. 2018. p. 214-218. 8492537 https://doi.org/10.1109/ICITE.2018.8492537