Summary: | 碩士 === 國立臺灣大學 === 電信工程學研究所 === 94 === In this thesis, we propose an architecture and algorithms for constructing a video cluster in a wireless multi-networking environment and implement a prototype system on a Linux platform. The proposed video cluster has three main features: (a) cost-effectiveness and distortion-controlled scheduling algorithm to distribute video frames over multiple networks, (b) adaptive video bit rate control according to the client’s network conditions,
(c) adaptive video distortion estimator to model various video sequences with different characteristics.
The goal of the proposed video cluster is to provide streaming video services with good quality and low cost. The challenge of providing such services is to take advantage of heterogeneous wireless network characteristics, including their bandwidth, packet
loss probability and transmission cost. To meet these requirements, we first present a detailed video distortion model to estimate the distortion experienced by video service users. Based on this model, we propose a 3-stage greedy algorithm, which determines
the transmission sequences over multiple heterogeneous wireless network channels, so that distortion controlled streaming video services can be provided with near-optimal cost.
Finally, we implement a proposed system and setup an experimental environment to verify the implementations, and video quality improved by our cost-effective scheduling algorithm. It is also shown that video bit rate can be adjusted according to varying
network conditions. At last, our video model may alter dynamically to fit different kinds of video sequences.
|