一個結合模糊邏輯控制與自適應田口方法的體驗品質預測模型應用於多重來源異質性行動網路

A QoE Prediction Model Combined Fuzzy Logic Control with Adaptive Taguchi Method in Multi-Source Heterogeneous Mobile Networks

黃品綸(Pin-Lun Huang)

指導教授 ： 陳永隆

國立臺中科技大學/資訊與流通學院/資訊工程系碩士班/碩士(2018年)

摘要

近年來，由於行動裝置和無線網路迅速發展以及進步，智慧型手機、平板電腦和筆記型電腦…等之行動裝置已成為了生活中不可缺少的一部分。而行動裝置的盛行，也帶動了無線網路媒體串流流量大量提升，根據Cisco的預測，到了2018年，行動網路影片流量將占總流量的75% [1]，爆炸性的流量增長使得行動影片串流服務在未來五年都將是最主要的行動網路應用之一，本論文將針對如何改善終端使用者串流影片體驗品質(Qualtiy of Experience，QoE)探討。
延續Priority-based transmission rate control with a fuzzy logical controller in wireless multimedia sensor networks中最高優先權的即時流量部分，其封包型態內容主要包含影片串流資訊，在此架構中，僅將流量做分配，達成網路資源的最大利用，未針對即時流量部分做完善的架構及設計，為改善行動網路串流體驗，我們引入體驗品質預測分配速率模型(QoE-Oriented Rate Allocation Scheme)，結合模糊邏輯控制(Fuzzy Logic Control，FLC)，提出QoE-Oriented Rate Allocation Multimedia Delivery Scheme with Fuzzy Logic Control (QRA-FLC)方法，將來源裝置電量、傳輸路徑連線品質及影片內容類型作為FLC輸入，輸出串流影片需求裝置最佳的傳輸速率。不過在上述架構中，所有影片都存放在影片伺服器，當同環境中需求的裝置數量增加，因為總頻寬為固定，傳輸速度和需求裝置數成反比與傳輸路徑可能互相干擾，串流體驗將大幅度的降低。為了解決此問題，進而我們加入了多重來源架構概念，提出Multi-Source QoE-Oriented Rate Allocation Multimedia Delivery Scheme with Fuzzy Logic Control (MSQRA-FLC)方法，讓欲串流收看某個影片的裝置，可以從同一網路拓樸架構中，其他已串流過該影片的多台裝置之間建立傳輸路徑，不但可以降低使用者等待的時間，也可避免其中一條串流傳輸路徑因干擾而影響串流體驗。最後，我們結合自適應田口方法(Adaptive Taguchi Method) 與FLC方法，提出Multi-Source QoE-Oriented Rate Allocation – Fuzzy Logic Control with Adaptive Taguchi Method (MSQRA-FLCAT)，利用機器學習概念使用電腦自我調配權重因子在不同條件下的最佳值，獲得每條路徑最佳的輸出串流速率，相較其他過去提出的方法，我們的方法能獲得最佳的體驗品質。

關鍵字

服務品質 ； 體驗品質 ； 多重來源 ； 模糊邏輯控制 ； 田口方法 ； 串流架構 ； 行動網路 ； 機器學習

並列摘要

In recent years, with the rapid development and progress of mobile devices and wireless network technologies, mobile devices such as smartphones, tablets and laptops have become an indispensable part of our lives. The prevalence of mobile devices has also led to a substantial increase in streaming media over the wireless network. According to Cisco's prediction, by 2018, mobile internet video traffic will account for 75% of the total traffic [1], and explosive traffic growth makes video streaming service will be one of the most important mobile network applications in the next five years. This thesis will discuss how to improve the quality of end-users’ streaming video quality of experience (QoE).
In priority-based transmission rate control with a fuzzy logical controller in wireless multimedia sensor networks, only the traffic is allocated to achieve the maximum utilization of network resources, it didn’t complete architecture and design for the immediate traffic part. To improve mobile network streaming experience, we combined QoE-oriented Scheme with fuzzy logic control (FLC), proposed QoE-Oriented Rate Allocation Multimedia Delivery Scheme with Fuzzy Logic control (QRA-FLC) method. This scheme uses demand device power, transmission connection quality and video content type as FLC input, furthermore, the FLC output is optimal transmission rate of each streaming video demand device. In this architecture, all videos are stored in the video source server. When the number of demand devices in the same environment increases, and the total bandwidth is fixed, users’ quality of experience will drop significantly because transmission speed is inversely proportional to the number of demand devices.
To solve this problem, we added the concept of multi-source architecture and proposed the new multi-source QoE-oriented rate allocation multimedia delivery scheme with fuzzy logic control method (MSQRA-FLC). This scheme allows demand devices be able to create a transmission path between the devices from other devices in the same network topology which have streamed the video. This scheme not only reduces users’ wait time, but also avoids streaming experience disrupted because of one of the transmission path interrupted.
Furthermore, we adopt adaptive Taguchi method combined with FLC, proposed multi-source QoE-oriented rate allocation – fuzzy logic control with adaptive Taguchi method (MSQRA-FLCAT). Our proposed MSQRA-FLCAT scheme exploits the concept of machine learning by setting the computer self-provision the best value of the weight factor under different conditions, we can obtain the optimal streaming rate for each path in multi-source transmission paths. Simulation results show that the quality of experience of our proposed MSQRA-FLCAT scheme outperforms the other schemes.

並列關鍵字

Quality of Service ； Quality of Experience ； Multi-Source Architecture ； Fuzzy Logic Control ； Taguchi Method ； Mobile Networks ； Streaming Path ； Machine Learning

參考文獻

[1] Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2016–2021, White Paper, 2017. [Online]. Available: https://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/mobile-white-paper-c11-520862.html

[2] L. Lei, Z. Zhong, C. Lin, and X. Shen, “Operator Controlled Device-to-Device Communications in LTE-Advanced Networks,” IEEE Wireless Communications, vol. 19, no. 3, pp. 96-104, Jun. 2012.

[3] C. Gao, A. Gutierrez, M. Rajan, R. Dreslinski, T. Mudge, and C. Wu, “A Study of Mobile Device Utilization,” in Proceedings of the IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS), Mar. 2015.

[4] B. A. Bjerke, “LTE-advanced and the evolution of LTE deployments,” IEEE Wireless Communications, vol. 18, no. 5, Oct. 2011.

[5] G. Larysa, S. Mariia, and S. Svitlana, “Method for resource allocation of virtualized network functions in hybrid environment,” in Proceedings of the IEEE International Black Sea Conference on Communications and Networking (BlackSeaCom), Jun. 2016.