CHOSUN

Artificial Intelligence for UAV-assisted 5G Heterogeneous NOMA Systems with Priority-based Joint Resource Allocation

Metadata Downloads
Author(s)
레즈완 시팟
Issued Date
2022
Abstract
5세대 무선 통신의 이기종 수요를 위해 massive machine type communication(mMTC), enhanced mobile broadband(eMBB), ultra reliable and low latency communication(URLLC) 기술이 발전하고 있다. 또한, 동일한 주파수에서 quality-of-service(QoS)를 보장하며 다수의 무선 신호를 동시 수신할 수 있도록 지원하기 위해, non-orthogonal multiple access(NOMA) 기술이 연구되고 있다. 이러한 NOMA 시스템에서 통신의 효율을 극대화하기 위해서는 최적의 자원 할당과 전송링크의 품질 보장이 해결되어야 한다.
본 논문에서는 5G 네트워크의 QoS 요구사항을 보장하고 네트워크 성능을 향상시키기 위해, 심층 Q학습 알고리즘을 사용하는 우선 순위 기반 채널할당 기술을 제안한다. NOMA 시스템의 다양한 제약조건을 통합하기 위해, Karush-Kuhn-Tucker(KKT) 조건을 활용하여 최적의 전력 할당 방식을 수학적으로 도출하고, 각 채널 및 전체 시스템의 sum-rate와 공정성을 최대화한다.
또한, UAV 지원 무선 네트워크를 고려하여, 연합학습을 활용한 강화학습 기반 다중 UAV-BS 탐색 방식을 제안한다. 이는 non-line-of-sight (NLOS), 열악한 무선 링크 품질 및 통신 범위, 다중 경로 페이딩에 의한 5G 네트워크 성능 저하를 극복하도록 설계한다.
다양한 시스템 매개변수에 대한 광범위한 시뮬레이션을 수행하여, 제안하는 기술이 다른 최신 기술에 비해 우수한 성능을 가짐을 확인한다.|For heterogeneous demands in fifth-generation (5G) new radio (NR), massive machine type communication (mMTC), enhanced mobile broadband (eMBB), and ultra-reliable and low-latency communication (URLLC) services have been introduced.
Non-orthogonal multiple access (NOMA) has been introduced to ensure these quality-of-services (QoS) requirements in which multiple devices can be served from the same frequency by manipulating the power domain and successive interference cancellation (SIC) technique.
To maximize the efficiency of NOMA systems, optimal resource allocation, ensuring transmission link quality are the key issues that need to be solved.
In this thesis, we propose a priority-based channel assignment with a deep $Q$-learning algorithm to maintain the 5G QoS requirements and increase the network performance. We formulate an optimal power allocation scheme under Karush–Kuhn–Tucker (KKT) optimality conditions incorporating different NOMA constraints. The main objectives are to maximize the channel sum-rate, system sum-rate, and system fairness.
We also propose a novel FDRL-based multiple UAV-BS navigation scheme to serve the 5G devices suffering from NLOS, poor link quality, and multi-path fading with maximum coverage, link quality, and fairness.
Finally, We conduct extensive simulations with respect to different system parameters and confirm that the proposed schemes perform better than other state-of-the-art schemes.
Alternative Title
UAV 지원 5G 이기종 NOMA 시스템의 우선순위 기반 자원할당을 위한 인공지능 기술 연구
Alternative Author(s)
Sifat Rezwan
Affiliation
조선대학교 일반대학원
Department
일반대학원 컴퓨터공학과
Advisor
최우열
Awarded Date
2022-02
Table Of Contents
ABSTRACT ⅵ

한글요약 ⅷ

I. INTRODUCTION 1
A.Related Works 3
1.Resource allocation 3
2.Link quality and Coverage 5
B.Contributions 8
C.Thesis Layout 10

II. Fundamentals of Reinforcement Learning 11
A.Reinforcement learning 11
B.Deep reinforcement learning 13
C.Federated deep reinforcement learning 16
1.HFDRL 17
2.VFDRL 19

III. Priority-based Joint Resource Allocation with DQL 22
A.Problem Statement 22
1.Multi-Carrier NOMA 22
2.System Model 23
3.Problem Formulation 26
B.Power Allocation 27
C.Priority-based Channel Assignment 29
1.Priority-based Channel Assignment 29
2.Deep Q-Learning Framework 30
3.Training 35
D.Simulation Analysis 37
1.Simulation Environment 39
2.Performance Analysis 40
3.Complexity and Parameter Analysis 47

IV. FDRL-Based Multi-UAV Navigation 51
A.Problem Statement 51
1.System Model 51
2.Problem Formulation 52
B.Proposed FDRL-based Multi-UAV Navigation 54
1.DeepQ-Learning Framework 54
2.Federated Learning Framework 57
3.Training 57
C.Simulation Analysis 59
1.Simulation Environment 59
2.Performance Analysis 62

PUBLICATIONS 74
D.Journals 74
E.Conferences 74

REFERENCES 75

ACKNOWLEDGEMENTS 87
Degree
Master
Publisher
조선대학교 대학원
Citation
레즈완 시팟. (2022). Artificial Intelligence for UAV-assisted 5G Heterogeneous NOMA Systems with Priority-based Joint Resource Allocation.
Type
Dissertation
URI
https://oak.chosun.ac.kr/handle/2020.oak/17240
http://chosun.dcollection.net/common/orgView/200000588743
Appears in Collections:
General Graduate School > 3. Theses(Master)
Authorize & License
  • AuthorizeOpen
  • Embargo2022-02-25
Files in This Item:

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.