以使用者行為預測法減少個人網路電視系統頻道切換時間之研究

Abstract

網路電視服務為一種利用寬頻IP網路作為介質而提供使用者電視服務之系統。有別於傳統的電視服務，網路協議電視服務受限於頻寬的限制，無法將所有頻道的數位內容廣播給使用者。因此當使用者決定從目前收看的頻道切換至另一個頻道時，必須重新的向伺服器要求此新頻道的串流內容(streaming content)。此時會有一段延遲的時間，稱為頻道切換延遲時間( channel zapping time)。通常，使用者能容忍的頻道切換延遲時間大約在兩秒左右。此外，由於近年來個人電腦的運算速度及穩定度不斷的提升，加上開放原始碼軟體社群的貢獻，使用者可以輕而易舉的建立一套專屬的網路協議電視服務跟親友分享。然而，在此環境下的頻道切換延遲時間仍是個亟需解決的問題。 為了解決這個問題，我們參考了許多在商業環境下減少頻道切換延遲時間的研究。先前大部分的研究都採用預先加入(prejoin)、預先緩衝( prebuffer )等機制在同時間內接受多個頻道的串流。在這篇碩士論文中，我們首先分析使用者轉台行為，並提出一個模型來描述此行為。藉由此分析，我們提出了一個預先緩衝演算法，此演算法可以精確的預測用戶的轉台行為，因此可有效降低所需的頻寬。在本碩士論文中，我們同時介紹我們自行建立的台大校園網路電視系統(NTU Campus IPTV System)，結合了隨選視訊播放(VOD)、YouTube等網路服務，提供使用者一個整合的、多元的家庭娛樂平台，提早實現未來互動式IP TV 家庭劇院之使用情境。

Internet protocol television (IPTV) is a system that provides various multimedia services based on broadband IP network. Unlike traditional television services, an IPTV system is always subject to bandwidth limitations, thus it can not broadcast all channels to subscribers at the same time. When a user decides to watch the channel from the current to another, he/she must re-request to the server for streaming content of this new channel. During this operation, there will be a time delay related to the channel change event, known as the channel zapping time. Due to the improvement of computational speed and stability of users’ PC in recent years, together with the contributions of open-source community, people can easily set up a personal IPTV system and share with their relatives and friends. However, the channel zapping time remains to be an important issue. To solve this problem, most previous studies have proposed to use prejoin (also named prebuffer in some documents) mechanism that accepts multiple channels at the same time. In this thesis, we first analyze the user behavior, and propose a model to describe this behavior. By this analysis, we proposed a prebuffer algorithm that can accurately predict the user behavior, and thus can effectively suppress the channel zapping time. We also introduce our NTU Campus IPTV system that combines the VOD, YouTube and other services. This system provides users an overview about what an integrated, interactive platform for home entertainment will be in the future.