Building a Lane Merge Coordination for ConnectedVehicles Using Deep Reinforcement Learning

Omar Nassef; Luis Sequeira Villarreal; Elias Salam; Toktam Mahmoodi

doi:10.1109/JIOT.2020.3017931

Building a Lane Merge Coordination for ConnectedVehicles Using Deep Reinforcement Learning

Omar Nassef, Luis Sequeira Villarreal, Elias Salam, Toktam Mahmoodi

Research output: Contribution to journal › Article › peer-review

22 Citations (Scopus)

118 Downloads (Pure)

Abstract

This paper presents a data-driven framework fortrajectory recommendation in automated and cooperative driv-ing. The considered cooperative driving manoeuvre is lane-mergecoordination, and while the trajectory recommendation can onlybe communicated to the connected vehicles, in computation ofthose recommendations both connected and unconnected vehiclesare taken into account. The data-driven framework is imple-mented centrally, comprising of two main components of aTrafficOrchestratorandData Fusion. TheTraffic Orchestratorpredictsthe safest trajectories for connected vehicles involved in the lane-merge manoeuvre. TheData Fusionincorporates camera detectedvehicles in order to map all vehicles including connected andunconnected. To this end, the recommendations are built usingvarious state-of-the-art machine learning techniques includingdeep reinforcement learning and dueling deep Q-network. Ourevaluations are conducted using the real-system deployed in thetest track, with a mix of connected and unconnected vehicles.The results demonstrate precision of predicted trajectories, andpercentage of successful lane merge achieved deploying differentmachine learning techniques.

Original language	English
Pages (from-to)	2540 - 2557
Journal	IEEE Internet of Things Journal
Volume	8
Issue number	4
DOIs	https://doi.org/10.1109/JIOT.2020.3017931
Publication status	Published - Feb 2021

Access to Document

10.1109/JIOT.2020.3017931

Lane_Merge_Coordination_LSTM_RL(accepted)Accepted author manuscript, 6.01 MB

https://ieeexplore.ieee.org/document/9171338

Cite this

@article{773467b8535d45abb4972751cbd69355,

title = "Building a Lane Merge Coordination for ConnectedVehicles Using Deep Reinforcement Learning",

abstract = "This paper presents a data-driven framework fortrajectory recommendation in automated and cooperative driv-ing. The considered cooperative driving manoeuvre is lane-mergecoordination, and while the trajectory recommendation can onlybe communicated to the connected vehicles, in computation ofthose recommendations both connected and unconnected vehiclesare taken into account. The data-driven framework is imple-mented centrally, comprising of two main components of aTrafficOrchestratorandData Fusion. TheTraffic Orchestratorpredictsthe safest trajectories for connected vehicles involved in the lane-merge manoeuvre. TheData Fusionincorporates camera detectedvehicles in order to map all vehicles including connected andunconnected. To this end, the recommendations are built usingvarious state-of-the-art machine learning techniques includingdeep reinforcement learning and dueling deep Q-network. Ourevaluations are conducted using the real-system deployed in thetest track, with a mix of connected and unconnected vehicles.The results demonstrate precision of predicted trajectories, andpercentage of successful lane merge achieved deploying differentmachine learning techniques.",

author = "Omar Nassef and {Sequeira Villarreal}, Luis and Elias Salam and Toktam Mahmoodi",

year = "2021",

month = feb,

doi = "10.1109/JIOT.2020.3017931",

language = "English",

volume = "8",

pages = "2540 -- 2557",

journal = "IEEE Internet of Things Journal",

issn = "2327-4662",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "4",

}

TY - JOUR

T1 - Building a Lane Merge Coordination for ConnectedVehicles Using Deep Reinforcement Learning

AU - Nassef, Omar

AU - Sequeira Villarreal, Luis

AU - Salam, Elias

AU - Mahmoodi, Toktam

PY - 2021/2

Y1 - 2021/2

N2 - This paper presents a data-driven framework fortrajectory recommendation in automated and cooperative driv-ing. The considered cooperative driving manoeuvre is lane-mergecoordination, and while the trajectory recommendation can onlybe communicated to the connected vehicles, in computation ofthose recommendations both connected and unconnected vehiclesare taken into account. The data-driven framework is imple-mented centrally, comprising of two main components of aTrafficOrchestratorandData Fusion. TheTraffic Orchestratorpredictsthe safest trajectories for connected vehicles involved in the lane-merge manoeuvre. TheData Fusionincorporates camera detectedvehicles in order to map all vehicles including connected andunconnected. To this end, the recommendations are built usingvarious state-of-the-art machine learning techniques includingdeep reinforcement learning and dueling deep Q-network. Ourevaluations are conducted using the real-system deployed in thetest track, with a mix of connected and unconnected vehicles.The results demonstrate precision of predicted trajectories, andpercentage of successful lane merge achieved deploying differentmachine learning techniques.

AB - This paper presents a data-driven framework fortrajectory recommendation in automated and cooperative driv-ing. The considered cooperative driving manoeuvre is lane-mergecoordination, and while the trajectory recommendation can onlybe communicated to the connected vehicles, in computation ofthose recommendations both connected and unconnected vehiclesare taken into account. The data-driven framework is imple-mented centrally, comprising of two main components of aTrafficOrchestratorandData Fusion. TheTraffic Orchestratorpredictsthe safest trajectories for connected vehicles involved in the lane-merge manoeuvre. TheData Fusionincorporates camera detectedvehicles in order to map all vehicles including connected andunconnected. To this end, the recommendations are built usingvarious state-of-the-art machine learning techniques includingdeep reinforcement learning and dueling deep Q-network. Ourevaluations are conducted using the real-system deployed in thetest track, with a mix of connected and unconnected vehicles.The results demonstrate precision of predicted trajectories, andpercentage of successful lane merge achieved deploying differentmachine learning techniques.

U2 - 10.1109/JIOT.2020.3017931

DO - 10.1109/JIOT.2020.3017931

M3 - Article

SN - 2327-4662

VL - 8

SP - 2540

EP - 2557

JO - IEEE Internet of Things Journal

JF - IEEE Internet of Things Journal

IS - 4

ER -

Building a Lane Merge Coordination for ConnectedVehicles Using Deep Reinforcement Learning

Abstract

Access to Document

Fingerprint

Cite this