A Deep Reinforcement Learning Agent for Traffic Intersection Control Optimization

Research output: Contribution to conferencePaper

Abstract

The efficiency of traffic flows in urban areas largely depends on signal operation. The state-of-the-art traffic signal control strategies are not able to efficiently deal with
varying or over-saturated conditions. To optimize the performance of existing traffic signal infrastructure, we present an end-to-end autonomous intersection control agent, based on Deep Reinforcement Learning (DRL). In the recent years, DRL
has emerged as a powerful tool, solving control problems involving sequential decision making and demonstrating unprecedented success in complex settings. Our DRL traffic intersection control agent configures the traffic signal regimes based solely
on live photo-realistic camera footage. We demonstrate that our agent consistently, significantly outperforms state-of-the-art fixed (pre-defined) and adaptive (induction loop-based) signal control methods under a wide range of ambient conditions, by
increasing the traffic throughput and decreasing the intersection traversal time for individual vehicles.
LanguageEnglish
Publication statusPublished - Oct 2019
Event22nd Intelligent Transportation Systems Conference, ITSC 2019 -
Duration: 27 Oct 201930 Oct 2019
Conference number: 22

Conference

Conference22nd Intelligent Transportation Systems Conference, ITSC 2019
Abbreviated titleITSC
Period27/10/1930/10/19

Fingerprint

Reinforcement learning
Traffic signals
Decision making
Cameras
Throughput

Cite this

Garg, D., Chli, M., & Vogiatzis, G. (2019). A Deep Reinforcement Learning Agent for Traffic Intersection Control Optimization. Paper presented at 22nd Intelligent Transportation Systems Conference, ITSC 2019, .
Garg, Deepeka ; Chli, Maria ; Vogiatzis, George. / A Deep Reinforcement Learning Agent for Traffic Intersection Control Optimization. Paper presented at 22nd Intelligent Transportation Systems Conference, ITSC 2019, .
@conference{b6734dce9d174645b2f5ddb58976d6ea,
title = "A Deep Reinforcement Learning Agent for Traffic Intersection Control Optimization",
abstract = "The efficiency of traffic flows in urban areas largely depends on signal operation. The state-of-the-art traffic signal control strategies are not able to efficiently deal withvarying or over-saturated conditions. To optimize the performance of existing traffic signal infrastructure, we present an end-to-end autonomous intersection control agent, based on Deep Reinforcement Learning (DRL). In the recent years, DRLhas emerged as a powerful tool, solving control problems involving sequential decision making and demonstrating unprecedented success in complex settings. Our DRL traffic intersection control agent configures the traffic signal regimes based solelyon live photo-realistic camera footage. We demonstrate that our agent consistently, significantly outperforms state-of-the-art fixed (pre-defined) and adaptive (induction loop-based) signal control methods under a wide range of ambient conditions, byincreasing the traffic throughput and decreasing the intersection traversal time for individual vehicles.",
author = "Deepeka Garg and Maria Chli and George Vogiatzis",
year = "2019",
month = "10",
language = "English",
note = "22nd Intelligent Transportation Systems Conference, ITSC 2019, ITSC ; Conference date: 27-10-2019 Through 30-10-2019",

}

Garg, D, Chli, M & Vogiatzis, G 2019, 'A Deep Reinforcement Learning Agent for Traffic Intersection Control Optimization' Paper presented at 22nd Intelligent Transportation Systems Conference, ITSC 2019, 27/10/19 - 30/10/19, .

A Deep Reinforcement Learning Agent for Traffic Intersection Control Optimization. / Garg, Deepeka; Chli, Maria; Vogiatzis, George.

2019. Paper presented at 22nd Intelligent Transportation Systems Conference, ITSC 2019, .

Research output: Contribution to conferencePaper

TY - CONF

T1 - A Deep Reinforcement Learning Agent for Traffic Intersection Control Optimization

AU - Garg, Deepeka

AU - Chli, Maria

AU - Vogiatzis, George

PY - 2019/10

Y1 - 2019/10

N2 - The efficiency of traffic flows in urban areas largely depends on signal operation. The state-of-the-art traffic signal control strategies are not able to efficiently deal withvarying or over-saturated conditions. To optimize the performance of existing traffic signal infrastructure, we present an end-to-end autonomous intersection control agent, based on Deep Reinforcement Learning (DRL). In the recent years, DRLhas emerged as a powerful tool, solving control problems involving sequential decision making and demonstrating unprecedented success in complex settings. Our DRL traffic intersection control agent configures the traffic signal regimes based solelyon live photo-realistic camera footage. We demonstrate that our agent consistently, significantly outperforms state-of-the-art fixed (pre-defined) and adaptive (induction loop-based) signal control methods under a wide range of ambient conditions, byincreasing the traffic throughput and decreasing the intersection traversal time for individual vehicles.

AB - The efficiency of traffic flows in urban areas largely depends on signal operation. The state-of-the-art traffic signal control strategies are not able to efficiently deal withvarying or over-saturated conditions. To optimize the performance of existing traffic signal infrastructure, we present an end-to-end autonomous intersection control agent, based on Deep Reinforcement Learning (DRL). In the recent years, DRLhas emerged as a powerful tool, solving control problems involving sequential decision making and demonstrating unprecedented success in complex settings. Our DRL traffic intersection control agent configures the traffic signal regimes based solelyon live photo-realistic camera footage. We demonstrate that our agent consistently, significantly outperforms state-of-the-art fixed (pre-defined) and adaptive (induction loop-based) signal control methods under a wide range of ambient conditions, byincreasing the traffic throughput and decreasing the intersection traversal time for individual vehicles.

M3 - Paper

ER -

Garg D, Chli M, Vogiatzis G. A Deep Reinforcement Learning Agent for Traffic Intersection Control Optimization. 2019. Paper presented at 22nd Intelligent Transportation Systems Conference, ITSC 2019, .