Skip to Main content Skip to Navigation
Theses

Coding with state information. Application to information embedding.

Abstract : The problem of coding with state information (CCSI) is a new interference cancellation technique for both data transmission and data compression. It concerns all the situations where the transmitter knows a part of the interference in the channel (via a feedback loop, for example). The goal is then to use this knowledge about the channel in order to conceive an efficient coding scheme. One potentially promising application, at the cross-road of both data transmission and data compression, is information embedding. The embedding of information poses many challenges in a variety of reseach areas. This involves information theory for assessing the theoretic limits of performance, signal processing for implementation issues and communication theory for code design. In this thesis, we consider the problem of information embedding in its three aspects. The work can be structured into four parts. In the first part, information embedding is mathematically formalized as a joint source-channel coding problem. For instance, we show that the problem of CCSI available at the transmitter is basically a joint source-channel coding probelem. Next, we use nested lattices for the design of good low-complexity algebraic-based codes. In the second part, we consider the problem of multiple user information embedding (recognized as a multi-user communication problem) and conceive structured codebooks and appropriate coding strategies that closely approache the theoretic limits. The third part concerns channel sensitivity to little perturbations of the state information. We evaluate the loss in performance due to a weak additive contaminating state information and show that, under certain circumstances, the transmitter must adapt to the available knowledge about the channel, by (eventually) changing its coding strategy. The fourth part determines the performance of CCSI over an AWGN channel with jitter (AWGN\&J) in a game theory context.
Document type :
Theses
Complete list of metadatas

https://pastel.archives-ouvertes.fr/pastel-00598336
Contributor : Abdellatif Zaidi <>
Submitted on : Monday, June 6, 2011 - 11:49:40 AM
Last modification on : Wednesday, September 16, 2020 - 4:42:28 PM
Long-term archiving on: : Sunday, December 4, 2016 - 11:26:46 PM

Identifiers

  • HAL Id : pastel-00598336, version 1

Collections

Citation

Abdellatif Zaidi. Coding with state information. Application to information embedding.. Electronics. Ecole nationale supérieure des telecommunications - ENST, 2005. English. ⟨pastel-00598336⟩

Share

Metrics

Record views

593

Files downloads

904