Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections

In recent years, the growing development of Connected Autonomous Vehicles (CAV), Intelligent Transport Systems (ITS), and 5G communication networks have led to the advent of Autonomous Intersection Management (AIM) systems. AIMs present a new paradigm for CAV control in future cities, taking control...

ver descrição completa

Detalhes bibliográficos
Autores: Guillén Pérez, Antonio, Cano Baños, María Dolores
Formato: artículo
Estado:Versión publicada
Fecha de publicación:2022
País:España
Recursos:Universidad Politécnica de Cartagena(UPCT)
Repositorio:Repositorio Digital UPCT
OAI Identifier:oai:repositorio.upct.es:10317/13773
Acesso em linha:http://hdl.handle.net/10317/13773
https://ieeexplore.ieee.org/document/9762548
Access Level:acceso abierto
Palavra-chave:Autonomous intersection management
Connected autonomous vehicles
Deep reinforcement learning
Intelligent transport systems
Intersection traffic management
Multi-agent deep reinforcement learning
Ingeniería Telemática
3317 Tecnología de Vehículos de Motor
id ES_de4dde2e90f2945050def90643af6d2e
oai_identifier_str oai:repositorio.upct.es:10317/13773
network_acronym_str ES
network_name_str España
repository_id_str
spelling Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersectionsGuillén Pérez, AntonioCano Baños, María DoloresAutonomous intersection managementConnected autonomous vehiclesDeep reinforcement learningIntelligent transport systemsIntersection traffic managementMulti-agent deep reinforcement learningIngeniería Telemática3317 Tecnología de Vehículos de MotorIn recent years, the growing development of Connected Autonomous Vehicles (CAV), Intelligent Transport Systems (ITS), and 5G communication networks have led to the advent of Autonomous Intersection Management (AIM) systems. AIMs present a new paradigm for CAV control in future cities, taking control of CAVs in scenarios where cooperation is necessary and allowing safe and efficient traffic flows, eliminating traffic signals. So far, the development of AIM algorithms has been based on basic control algorithms, without the ability to adapt or keep learning new situations. To solve this, in this paper we present a new advanced AIM approach based on end-to-end Multi-Agent Deep Reinforcement Learning (MADRL) and trained using Curriculum through Self-Play, called advanced Reinforced AIM (adv.RAIM). adv.RAIM enables the control of CAVs at intersections in a collaborative way, autonomously learning complex real-life traffic dynamics. In addition, adv.RAIM provides a new way to build smarter AIMs capable of proactively controlling CAVs in other highly complex scenarios. Results show remarkable improvements when compared to traffic light control techniques (reducing travel time by 59% or reducing time lost due to congestion by 95%), as well as outperforming other recently proposed AIMs (reducing waiting time by 56%), highlighting the advantages of using MADRL.This work was supported in part by MCIN/AEI/10.13039/501100011033 under Grant PID2020-116329GB-C22, and in part by the Fundación Séneca, Región de Murcia, Spain under Grant 20740/FPI/18.IEEEMinisterio de Ciencia e InnovaciónFundación Séneca202420242022info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionapplication/pdfapplication/pdfhttp://hdl.handle.net/10317/13773https://ieeexplore.ieee.org/document/9762548reponame:Repositorio Digital UPCTinstname:Universidad Politécnica de Cartagena(UPCT)Inglésinfo:eu-repo/grantAgreement/AEI/Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020/PID2020-116329GB-C22PID2020-116329GB-C2220740/FPI/18info:eu-repo/semantics/openAccessoai:repositorio.upct.es:10317/137732026-05-15T06:39:02Z
dc.title.none.fl_str_mv Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections
title Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections
spellingShingle Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections
Guillén Pérez, Antonio
Autonomous intersection management
Connected autonomous vehicles
Deep reinforcement learning
Intelligent transport systems
Intersection traffic management
Multi-agent deep reinforcement learning
Ingeniería Telemática
3317 Tecnología de Vehículos de Motor
title_short Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections
title_full Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections
title_fullStr Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections
title_full_unstemmed Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections
title_sort Multi-agent deep reinforcement learning to manage connected utonomous vehicles at tomorrow's intersections
dc.creator.none.fl_str_mv Guillén Pérez, Antonio
Cano Baños, María Dolores
author Guillén Pérez, Antonio
author_facet Guillén Pérez, Antonio
Cano Baños, María Dolores
author_role author
author2 Cano Baños, María Dolores
author2_role author
dc.contributor.none.fl_str_mv Ministerio de Ciencia e Innovación
Fundación Séneca
dc.subject.none.fl_str_mv Autonomous intersection management
Connected autonomous vehicles
Deep reinforcement learning
Intelligent transport systems
Intersection traffic management
Multi-agent deep reinforcement learning
Ingeniería Telemática
3317 Tecnología de Vehículos de Motor
topic Autonomous intersection management
Connected autonomous vehicles
Deep reinforcement learning
Intelligent transport systems
Intersection traffic management
Multi-agent deep reinforcement learning
Ingeniería Telemática
3317 Tecnología de Vehículos de Motor
description In recent years, the growing development of Connected Autonomous Vehicles (CAV), Intelligent Transport Systems (ITS), and 5G communication networks have led to the advent of Autonomous Intersection Management (AIM) systems. AIMs present a new paradigm for CAV control in future cities, taking control of CAVs in scenarios where cooperation is necessary and allowing safe and efficient traffic flows, eliminating traffic signals. So far, the development of AIM algorithms has been based on basic control algorithms, without the ability to adapt or keep learning new situations. To solve this, in this paper we present a new advanced AIM approach based on end-to-end Multi-Agent Deep Reinforcement Learning (MADRL) and trained using Curriculum through Self-Play, called advanced Reinforced AIM (adv.RAIM). adv.RAIM enables the control of CAVs at intersections in a collaborative way, autonomously learning complex real-life traffic dynamics. In addition, adv.RAIM provides a new way to build smarter AIMs capable of proactively controlling CAVs in other highly complex scenarios. Results show remarkable improvements when compared to traffic light control techniques (reducing travel time by 59% or reducing time lost due to congestion by 95%), as well as outperforming other recently proposed AIMs (reducing waiting time by 56%), highlighting the advantages of using MADRL.
publishDate 2022
dc.date.none.fl_str_mv 2022
2024
2024
dc.type.none.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.none.fl_str_mv http://hdl.handle.net/10317/13773
https://ieeexplore.ieee.org/document/9762548
url http://hdl.handle.net/10317/13773
https://ieeexplore.ieee.org/document/9762548
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.relation.none.fl_str_mv info:eu-repo/grantAgreement/AEI/Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020/PID2020-116329GB-C22
PID2020-116329GB-C22
20740/FPI/18
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
application/pdf
dc.publisher.none.fl_str_mv IEEE
publisher.none.fl_str_mv IEEE
dc.source.none.fl_str_mv reponame:Repositorio Digital UPCT
instname:Universidad Politécnica de Cartagena(UPCT)
instname_str Universidad Politécnica de Cartagena(UPCT)
reponame_str Repositorio Digital UPCT
collection Repositorio Digital UPCT
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869421961792192512
score 15.811543

Registos relacionados