Distributed Source Coding: Theory, Algorithms, and Applications

点击翻译按钮获取中文摘要

Abstract for the project:Distributed source coding (DSC) refers to the compression of the outputsof two or more physically separated sources that do not communicate witheach other. Driven by a host of emerging applications (e.g., distributedsensor networks and wireless video), DSC has assumed renewed interest inrecent years. Although the theoretical result given by the Slepian-Wolftheorem has been known for more than 30 years, practical approaches toDSC did not appear until 1999.Aiming to advance the field, this research covers the theory, algorithms,and applications of DSC, with focus on the hard problem of practical codedesign. The basic element of a distributed source code is the binningscheme. For Wyner-Ziv code design, the PI proposes a paradigm based onSlepian-Wolf coded quantization (SWCQ). Preliminary analytical resultsshow that the performance limits of SWCQ mirror those in classic sourcecoding. Initial code design based on TCQ and LDPC codes gives results thatare significantly better than any previously reported. Inspired by theseresults, the PI proposes to study successive refinement for the Wyner-Zivproblem and apply SWCQ to layered video coding.This research will also tackle the more general problem of multiterminalsource coding under the same powerful paradigm of SWCQ. While asymmetricSlepian-Wolf coding is used in SWCQ for Wyner-Ziv coding, formultiterminal source coding with SWCQ, the PI will explore symmetricSlepian-Wolf coding after quantization of each source. Both directmultiterminal source coding and indirect multiterminal source coding (e.g., the CEO problem that arises in distributed sensor networks) will be treated.The theme of the research is to view lossy DSC as a source-channel codingproblem and perform algebraic binning via the combination of quantizationand Slepian-Wolf coding based on channel codes. The main novelty lies inusing a single SWCQ framework as the embodiment of algebraic binning tosolve different code design problems and address potential applications.