Veröffentlichungen – Details

Iterative Source-Channel Decoding for Digital Mobile Communications

Autor:

Marc Adrat

Herausgeber:

Peter Vary

Typ:

Dissertation

Reihe:

Aachener Beiträge zu Digitalen Nachrichtensystemen (ABDN)

Nummer:

16

Lehreinrichtung:

IND, RWTH Aachen

Verlag:

Verlag Mainz

Veröffentlichungsort:

Aachen, Germany

Veröffentlich.datum:

Juli 2003

ISBN:

3-860-73835-6

Sprache:

Englisch

Kurzfassung

In mobile communications, speech, audio, or video signals are typically encoded using two separate processing steps. Source encoding shall eliminate redundancy and irrelevance from the source signal, and channel encoding shall re-introduce artificial redundancy such that error correction becomes possible. In practice, however, some natural residual source redundancy remains after source encoding and some residual bit errors can appear after channel decoding.

The error robustness of communication systems can be improved by joint source-channel (de-)coding as discussed in this thesis. At the receiver, the reliability gains resulting from both, the artificial redundancy (channel coding) as well as the residual natural redundancy (source coding), are evaluated iteratively in a TURBO-like process. After the specific reliability gain due to the residual natural redundancy has carefully been analyzed, the general convergence behavior of the iterative process is studied using EXIT charts. The error concealing capabilities are demonstrated by computer simulation using realistic constraints. Powerful algorithms are derived for the efficient, nearly-optimal utilization of the residual natural redundancy.

Compared to conventional non-iterative transmission schemes, significant improvements of error robustness are achieved by iterative source-channel decoding.