Please use this identifier to cite or link to this item:
Title: 異質性網路中Wi-Fi與LAA共存分析建模與基於人工智慧方法公平性分配無線資源之研究
Analytical Modeling of Heterogeneous Network Wi-Fi and LTE LAA Coexistence Throughput and Using Artificial Intelligence Method for Fairness Allocation Radio Resources
Authors: 王嘉斌
Wang, Chia-pin
Fan Chiang, Tsung-Yi
Keywords: 微小型基地台
IEEE 802.11
Small cell
Unlicensed band
IEEE 802.11
Reinforcement Learning
Resource allocation
Issue Date: 2020
Abstract: 在即將到來的5G新無線電(New Radio, NR) 時代中,使用未授權頻段增加傳輸速度已是未來的趨勢,NR微小型基地台(small cell) 和Wi-Fi 將會被布建在同一場域或是整合至同一台機器中,並爭奪未授權頻段的使用權。由3GPP 主導的技術中LTE 許可輔助存取(Licensed Assisted Access, LAA),導入了LBT 技術,目的在於可以更加公平的競爭,使得Wi-Fi 與LAA 可在同一種場域共存。 本論文將介紹一個評估Wi-Fi和LAA共存下吞吐量 (Throughput)的框架,分別比較DCF 與LBT 各自的特徵,修改自著名的Bianchi 模型,並加入各種網路參數,如媒體存取控制層(MAC層)物理層(PHY)通道條件等進行模擬分析,並將呈現數值分析結果。 但是,我們發現在不同802.11協定下,LAA 與Wi-Fi 競爭會讓吞吐量出現極度不平衡的關係,為了改善不平等的關係,我們在本篇論文中提出基於強化式學習下調整TXOP (Transmission Opportunity)時間,來改善次世代異質性網路得公平效能,這項技術可以安裝在微小型基地台,增進在不同異質性網路下達到公平且有效的資源分配,保護網路用戶的服務品質。
In the coming era of 5G New Radio (NR), using unlicensed frequency bands to increase the transmission speed is the future trend. NR small cells and Wi-Fi will be deployed in the same field or integrated into the same machine and contend for the right to use unlicensed bands. In the technology led by 3GPP, LTE Licensed Assisted Access (LAA), the introduction of Listen Before Talk (LBT), aims at a fairer competition. so that Wi-Fi and LAA can coexist in the same field. This thesis will introduce a framework of evaluating the throughput under the coexistence of Wi-Fi and LAA. Compare the characteristics of DCF and LBT respectively, modify from the famous Bianchi model, and add various network parameters, such as Media Access Control (MAC) and physical layer (PHY) channel conditions, etc. For simulation and numerical analysis. However, we found that under different 802.11 protocols, the contention between LAA and Wi-Fi will cause an extremely uneven relationship in throughput. In order to improve the unequal relationship, we propose in this paper to adjust the Transmission Opportunity (TXOP) duration based on reinforcement learning to improve the fair performance of next-generation heterogeneous networks. This research can be installed in small cell to improve the fair and effective resource allocation protect the services of each network user’s quality.
Other Identifiers: G060775003H
Appears in Collections:學位論文

Files in This Item:
There are no files associated with this item.

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