運用強化學習與模糊理論於叢集式車載網路及其異質網路存取方法之研究;Study of Cluster-based Vehicular Network for Heterogeneous Network Access by Using Reinforcement Learning and Fuzzy Theory

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Communication Engineering > Electronic Thesis & Dissertation > Item 987654321/80971

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/80971

Title:	運用強化學習與模糊理論於叢集式車載網路及其異質網路存取方法之研究;Study of Cluster-based Vehicular Network for Heterogeneous Network Access by Using Reinforcement Learning and Fuzzy Theory
Authors:	陳昱伶;Chen, Yu-Ling
Contributors:	通訊工程學系
Keywords:	車聯網;IEEE 802.11p;C-V2X;機器學習;模糊理論;Vehicular communication;IEEE 802.11p;C-V2X;Machine Learning;Fuzzy Theory
Date:	2019-07-23
Issue Date:	2019-09-03 15:22:56 (UTC+8)
Publisher:	國立中央大學
Abstract:	汽車移動物聯網技術(簡稱：車聯網)是因應自動駕駛而發展的重要技術，目前第三代合作夥伴計劃(the Third Generation Partnership Project，3GPP)與Wi-Fi聯盟(Wi-Fi Alliance，WFA)兩大通訊標準組織正積極研究相關技術。由於兩組織所訂定的標準擁有各自的特色，再加上車輛高速移動與拓撲快速變化的特性，在各自傳輸資源有限的狀態下，車輛該如何抉擇通訊界面將是未來需要解決的問題。然而將車輛分散於兩種通訊界面中，傳輸資源也會因為每一輛車與中心設備(AP、Evolved Node B(eNodeB)…等)建立連線所需的控制訊息而佔據多數資源，因此該如何設計資料傳輸路徑，使得有些車輛可以不用與中心設備建立連線，而是透過鄰近車輛幫忙轉送，以降低車輛與中心設備所建立的連線數將是重要議題之一。本論文所設計的演算法根據系統效能與車輛所給予的資訊，判斷車輛該選擇何種通訊介面，以及如何將車輛分群，以提高資料傳送成功率與有效降低連線數。 ;Vehicular communication is a technology developed to support autonomous driving. Recently, the two major communication standards organization, Third Generation Partnership Project (3GPP) and Wi-Fi Alliance (WFA), are actively studying related technologies. However, each of the two standards has its own characteristics, along with some features such as high mobility of vehicles, frequent topology changes and limited transmission resources. Consequently, how to select the communication interface will be the problem that we should solve in the future. However, even if we disperse the vehicles into two different communication interface, the control signaling overhead caused by the connections between vehicles and the central equipment will exhaust the transmission resources. Therefore, the issue of designing a proper transmission model will be important. In other words, we aim to reduce the number of connections between the vehicle and the central equipment, achieved by forcing the vehicle to transmit data via its neighbor. According to information given by vehicles and environment condition of the system, we proposed an algorithm that decides which communication interface the vehicle should choose, as well as a vehicle clustering method, in order to improve packet delivery ratio and reduce the number of connections efficiently.
Appears in Collections:	[Graduate Institute of Communication Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	225	View/Open

社群 sharing

Loading...