Date of Award

6-2009

Degree Type

Thesis

Degree Name

Master of Applied Science (MASc)

Department

Electrical and Computer Engineering

Supervisor

Sorina Dumitrescu

Co-Supervisor

Shahram Shirani

Language

English

Abstract

This thesis presents a novel unequal erasure protection (UEP) strategy for the transmission of scalable data, formed by interleaving independently decodable and scalable streams, over packet erasure networks. The technique, termed multi-stream unequal erasure protection (M-UEP) differs from UEP by placing separate streams in separate packets to establish independence and using permuted systematic Reed-Solomon codes to enhance the distribution of message symbols amongst the packets. M-UEP improves upon UEP by ensuring that all received source symbols are decoded. The R-D optimal redundancy allocation problem for M-UEP is formulated and its globally optimal solution is shown to have a time complexity of O(2^N N(L+1)^(N+1)), where N is the number of packets and L is the packet length. To address the high complexity of the globally optimal solution, an efficient sub-optimal algorithm which runs in O(N^2 L^2) time is formulated. The additional side information necessary for M-UEP at the decoder is discussed and an upper bound on the amount of side information is derived. To mitigate the necessary side information, a technique termed (FM-UEP) is presented. A constrained optimal algorithm for generating N substreams from P primary substreams (P > N) is formulated, where the placement of primary substreams into N groups is constrained to a fixed-order. Four possible fixed-orders are proposed: raster scan, zig-zag scan, dispersed dot dithered and subband dispersed. Experiments performed on SPIRT coded images (with appropriate grouping of wavelet coefficient) validate the superiority of M-UEP and FM-UEP over UEP, with peak improvements of 0.6 and 0.5 dB, respectively. Additionally, our tests reveal that M-UEP is more robust than UEP in adverse, unpredictable and varying channel conditions.

McMaster University Library

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS