周承復Chou, Cheng-Fu臺灣大學:電機工程學研究所劉維倫Liu, Wei-LunWei-LunLiu2010-07-012018-07-062010-07-012018-07-062009U0001-0508200917582800http://ntur.lib.ntu.edu.tw//handle/246246/188158在無線網路環境下,多重傳送時考慮每個接收者的網路狀況並且定的做傳輸速率的調節是愈來愈困難的課題.首先碰到的問題是同時接多位接收者回報的訊息會在發送端發生碰撞,若使用分時多工的方法有大量接收者時會造成時間的浪費.之前的方法提出選一位網路狀況差的接收者當代表負責回報訊息,由此調整傳輸速率藉以保證網路上所有接收者都能接收到資料. 然而這些方法雖然保證每個接收者都能收到資料,卻處罰了處在網路狀況比較好能接收到高傳輸效率的一群收者. 因此在這篇論文中,我們提出一個分群式的異質性網路傳輸速調節機制讓每位接收者能夠依據他們所在的網路狀況接收自己能收最好的影片品質. 我們的方法會使用分群式的架構, 對每一群不同能的接收者選出一位代表者, 由這些代表者來預估整體接收者接收影片質的情形. 接者每張影片選擇最好的傳輸速率來達成最大化網路所有收者的收看品質. 最後由實驗能夠發現我們的方法會比之前找最差的收者當代表的方法平均收看的峰值信噪比(PSNR)好2 ⣵76; 5dB,且也能適應不同的網路狀況與不同的影片可變碼率(video bit rate).Rate adaptation for video multicast over wireless networks is a challengingroblem because of network dynamics, variable video bit rates, andeterogeneous clients who may expect differentiated video qualities. Priororks on the leader-based scheme select the channel rate that can provideeliable transmission for the node who experiences the worst channel condition.owever, this may penalize other nodes that can achieve a higherhroughput by receiving at a higher rate. Hence, in this work, we investigate rate-adaptive video multicast scheme that can provide heterogeneouslients differentiated visual qualities matching their channel conditions. Weirst propose a rate scheduling model that can select an optimal transmissionit-rate for each video frame in order to maximize the total visual qualityor a multicast group subject to the minimum-visual-quality-guaranteed constraint.e then present a practical protocol, called CRAM, which constructs cluster-based structure to characterize node heterogeneity and adapts thehannel rate based on video quality perceived by the representative clustereads. Since CRAM selects the rate by a sample-based technique, it is suitableor real-time video streaming without the need of any pre-process. Theimulation results show that CRAM can produce a gain of 2-5dB in terms ofhe average video quality as compared with the leader-based approach, anddapt to network dynamics and variable video bit rates.口試委員會審定書i謝iii文摘要vbstract vii Introduction 1 Related Work 5 Rate Scheduling Model 7 CRAM Framework 11.1 Cluster Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 PSNR-guaranteed Quality-differentiated Rate Selection . . . . . . . . . . 13.3 Two-stage Rate Adaptation . . . . . . . . . . . . . . . . . . . . . . . . . 15 Performance Evaluation 17.1 Video Quality Comparison . . . . . . . . . . . . . . . . . . . . . . . . . 18.2 Impact of Rate Adaptation . . . . . . . . . . . . . . . . . . . . . . . . . 20.3 Impact of Two-State Adaptation . . . . . . . . . . . . . . . . . . . . . . 22 Conclusions 23ibliography 25973205 bytesapplication/pdfen-US即時傳輸異質網路傳輸速率調節real-time streamingheterogeneous environmentrate adaptationthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/188158/1/ntu-98-R96922061-1.pdf