Modeling and Comparison of Video Coding Techniques for Video Streaming in Wireless Mesh Networks: A Network Optimization Perspective
Date Issued
2008
Date
2008
Author(s)
Tsai, Tsung-Lin
Abstract
The objective of this thesis is to analyze and optimize video performance using different coding techniques over wireless mesh networks. Most work focuses on achieving the highest throughput for performance optimization in mesh networks. However, this goal is not always the best for video streaming, and the characteristics of video coding techniques need to be considered. In this thesis, we first investigate the impact of network factors on optimal video quality. We build an optimization framework for video streaming in mesh networks by considering rate-distortion optimization of different coding techniques. The optimization problem can be formulated as a mixed integer nonlinear programming, and we evaluate the performance of different coding techniques, including MPEG-4, Fine Granularity Scalability (FGS), and Multiple Description Coding (MDC) in the proposed framework. We find that although MDC has poor coding efficiency, it performs better than MPEG-4 and FGS when the link capacity is small. We next extend the proposed framework to incorporate the impact of routing layer mechanisms, including minimum delay routing, minimum hop routing, and minimum distance routing. We discuss their difference from the optimal routing, and find that minimum delay routing performs best, especially when MDC is used. Finally, we extend the framework to incorporate the impact of MAC layer mechanisms, including transmission power control and rate adaptation. We find that among the different coding techniques, FGS improves most when using MAC layer mechanisms. MDC has stable performance for different network conditions due to the independency of multiple descriptions. We conclude that the optimal choice of video coding techniques should consider the routing and MAC layer mechanisms in mesh networks.
Subjects
Video Coding Techniques
Mesh Networks
Optimization
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-97-R95942101-1.pdf
Size
23.32 KB
Format
Adobe PDF
Checksum
(MD5):d67fd26ee0996ed0f00713217d3972f3