LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning

Luke Maguire; Mien Van; Kabirat Olayemi; Seán McLoone

doi:10.1109/ICM62621.2025.10934926

LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning

School of Electronics, Electrical Engineering and Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

59 Downloads (Pure)

Abstract

A novel deep reinforcement learning (DRL) framework is proposed for autonomous navigation of unmanned ground vehicles (UGVs) in dynamic environments using LiDAR data. Utilizing a Graph Attention Network (GAT), the framework processes high-dimensional LiDAR data into a meaningful state representation, enabling informed navigation decisions and adaptability to changing obstacles. The approach is validated through simulations of a Husky A200 UGV, demonstrating significant performance improvements over baseline TD3 algorithms. Key results include an improved success rate in trained environments and enhanced generalization in unseen scenarios, with reduced collisions and shorter navigation times. The findings highlight the potential of GAT-enhanced DRL for efficient and cost-effective autonomous navigation in real-world applications.

Original language	English
Title of host publication	2025 IEEE International Conference on Mechatronics (ICM’25): Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Number of pages	6
ISBN (Electronic)	9798331533892
ISBN (Print)	9798331533908
DOIs	https://doi.org/10.1109/ICM62621.2025.10934926
Publication status	Published - 26 Mar 2025
Event	2025 IEEE International Conference on Mechatronics (ICM’25) - Wollongong, Australia Duration: 28 Feb 2025 → 02 Mar 2025

Publication series

Name	IEEE International Conference on Mechatronics (ICM): Proceedings
ISSN (Print)	2837-1143
ISSN (Electronic)	2837-1151

Conference

Conference	2025 IEEE International Conference on Mechatronics (ICM’25)
Country/Territory	Australia
City	Wollongong
Period	28/02/2025 → 02/03/2025

Publications and Copyright Policy

This work is licensed under Queen’s Research Publications and Copyright Policy.

Keywords

LiDAR
LiDAR-only navigation
UGVs
graph attention networks

Access to Document

10.1109/ICM62621.2025.10934926Licence: Unspecified

LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning
Copyright 2025 the authors. This is an accepted manuscript distributed under a Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited.
Accepted author manuscript, 1.16 MBLicence: CC BY

Cite this

Maguire, L., Van, M., Olayemi, K., & McLoone, S. (2025). LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning. In 2025 IEEE International Conference on Mechatronics (ICM’25): Proceedings (IEEE International Conference on Mechatronics (ICM): Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICM62621.2025.10934926

Maguire, Luke ; Van, Mien ; Olayemi, Kabirat et al. / LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning. 2025 IEEE International Conference on Mechatronics (ICM’25): Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. (IEEE International Conference on Mechatronics (ICM): Proceedings).

@inproceedings{094155d7c2d048dc920f9ad2aae812a0,

title = "LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning",

abstract = "A novel deep reinforcement learning (DRL) framework is proposed for autonomous navigation of unmanned ground vehicles (UGVs) in dynamic environments using LiDAR data. Utilizing a Graph Attention Network (GAT), the framework processes high-dimensional LiDAR data into a meaningful state representation, enabling informed navigation decisions and adaptability to changing obstacles. The approach is validated through simulations of a Husky A200 UGV, demonstrating significant performance improvements over baseline TD3 algorithms. Key results include an improved success rate in trained environments and enhanced generalization in unseen scenarios, with reduced collisions and shorter navigation times. The findings highlight the potential of GAT-enhanced DRL for efficient and cost-effective autonomous navigation in real-world applications.",

keywords = "LiDAR, LiDAR-only navigation, UGVs, graph attention networks",

author = "Luke Maguire and Mien Van and Kabirat Olayemi and Se{\'a}n McLoone",

year = "2025",

month = mar,

day = "26",

doi = "10.1109/ICM62621.2025.10934926",

language = "English",

isbn = "9798331533908",

series = "IEEE International Conference on Mechatronics (ICM): Proceedings",

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

booktitle = "2025 IEEE International Conference on Mechatronics (ICM{\textquoteright}25): Proceedings",

address = "United States",

note = "2025 IEEE International Conference on Mechatronics (ICM{\textquoteright}25) ; Conference date: 28-02-2025 Through 02-03-2025",

}

Maguire, L , Van, M , Olayemi, K & McLoone, S 2025, LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning. in 2025 IEEE International Conference on Mechatronics (ICM’25): Proceedings. IEEE International Conference on Mechatronics (ICM): Proceedings, Institute of Electrical and Electronics Engineers Inc., 2025 IEEE International Conference on Mechatronics (ICM’25) , Wollongong, Australia, 28/02/2025. https://doi.org/10.1109/ICM62621.2025.10934926

LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning. / Maguire, Luke ; Van, Mien ; Olayemi, Kabirat et al.
2025 IEEE International Conference on Mechatronics (ICM’25): Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. (IEEE International Conference on Mechatronics (ICM): Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning

AU - Maguire, Luke

AU - Van, Mien

AU - Olayemi, Kabirat

AU - McLoone, Seán

PY - 2025/3/26

Y1 - 2025/3/26

N2 - A novel deep reinforcement learning (DRL) framework is proposed for autonomous navigation of unmanned ground vehicles (UGVs) in dynamic environments using LiDAR data. Utilizing a Graph Attention Network (GAT), the framework processes high-dimensional LiDAR data into a meaningful state representation, enabling informed navigation decisions and adaptability to changing obstacles. The approach is validated through simulations of a Husky A200 UGV, demonstrating significant performance improvements over baseline TD3 algorithms. Key results include an improved success rate in trained environments and enhanced generalization in unseen scenarios, with reduced collisions and shorter navigation times. The findings highlight the potential of GAT-enhanced DRL for efficient and cost-effective autonomous navigation in real-world applications.

AB - A novel deep reinforcement learning (DRL) framework is proposed for autonomous navigation of unmanned ground vehicles (UGVs) in dynamic environments using LiDAR data. Utilizing a Graph Attention Network (GAT), the framework processes high-dimensional LiDAR data into a meaningful state representation, enabling informed navigation decisions and adaptability to changing obstacles. The approach is validated through simulations of a Husky A200 UGV, demonstrating significant performance improvements over baseline TD3 algorithms. Key results include an improved success rate in trained environments and enhanced generalization in unseen scenarios, with reduced collisions and shorter navigation times. The findings highlight the potential of GAT-enhanced DRL for efficient and cost-effective autonomous navigation in real-world applications.

KW - LiDAR

KW - LiDAR-only navigation

KW - UGVs

KW - graph attention networks

U2 - 10.1109/ICM62621.2025.10934926

DO - 10.1109/ICM62621.2025.10934926

M3 - Conference contribution

SN - 9798331533908

T3 - IEEE International Conference on Mechatronics (ICM): Proceedings

BT - 2025 IEEE International Conference on Mechatronics (ICM’25): Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2025 IEEE International Conference on Mechatronics (ICM’25)

Y2 - 28 February 2025 through 2 March 2025

ER -

Maguire L , Van M , Olayemi K , McLoone S. LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning. In 2025 IEEE International Conference on Mechatronics (ICM’25): Proceedings. Institute of Electrical and Electronics Engineers Inc. 2025. (IEEE International Conference on Mechatronics (ICM): Proceedings). doi: 10.1109/ICM62621.2025.10934926

LiDAR-only navigation of UGVs in dynamic environments via graph attention networks and deep reinforcement learning

Abstract

Publication series

Conference

Publications and Copyright Policy

Keywords

Access to Document

Fingerprint

Cite this