A4 Article in conference proceedings
Multi-agent Deep Reinforcement Learning-based Trajectory Design for UAV-aided Edge Computing System (2023)

Lu, G., & Chang, Z. (2023). Multi-agent Deep Reinforcement Learning-based Trajectory Design for UAV-aided Edge Computing System. In WCNC 2023 : Proceedings of the IEEE Wireless Communications and Networking Conference. IEEE. IEEE Wireless Communications and Networking Conference. https://doi.org/10.1109/wcnc55385.2023.10118895

JYU authors or editors

Publication details

All authors or editors: Lu, Gengyuan; Chang, Zheng

Parent publication: WCNC 2023 : Proceedings of the IEEE Wireless Communications and Networking Conference

Place and date of conference: Glasgow, United Kingdom, 26.-29.3.2023

ISBN: 978-1-6654-9123-5

eISBN: 978-1-6654-9122-8

Journal or series: IEEE Wireless Communications and Networking Conference

ISSN: 1525-3511

eISSN: 1558-2612

Publication year: 2023

Publisher: IEEE

Publication country: United States

Publication language: English

DOI: https://doi.org/10.1109/wcnc55385.2023.10118895

Publication open access: Not open

Publication channel open access: 

Abstract

The unmanned aerial vehicle (UAV)-aided communication system is able to extend the coverage of the cellular network and provide flexible connectivity to the ground users. Recently, in order to achieve the ambitious goal of ubiquitous computing and intelligence, UAV-aided mobile edge computing (MEC) system also attracts significant research interests. In this work, we investigate the trajectory design for a multi-UAV-aided MEC system to maximize the defined quality of experience (QoE) and service fairness of multiple ground user equipments (UEs). Since the practical communication environment is complex and time-varying, it is challenging for the multiple UAVs to dynamically make autonomous decisions. Meanwhile, the centralized decision-making scheme may also have a certain degree of transmission delay and affect efficiency of the considered system. Therefore, bearing in mind these challenges, in order to solve the formulated problem, we propose a multi-agent deep reinforcement learning algorithm using Multi-Agent Deep Deterministic Policy Gradient (MADDPG) to design UAV’s trajectory via optimizing the flying angle and distance. The simulation results show that our proposed solution outperforms the traditional algorithms.

Keywords: unmanned aerial vehicles; information and communications technology; communications technology; data systems

Free keywords: mobile edge computing; unmanned aerial vehicle; multi-agent deep reinforcement learning; trajectory design

Contributing organizations

Ministry reporting: Yes

Reporting Year: 2023

JUFO rating: 1