Student:

Robin Nyombi Schofield Ssenyonga

Advisor:
Prof. Dr.-Ing. Werner Henkel

Time frame:
Spring 2015

 

 

 

 

 

Description:

Neural network rules and coding theory concepts have been associated in the past. It is known that in relation to neural networks, iterative decoding exhibits some nonlinear dynamics. The purpose of this thesis, therefore, is to set forth an evaluation of the concept of neural network design related to Low-Density Parity-Check (LDPC) code structures and algorithmic aspects. A basic introduction to neural networks is given as major aspects of existing network structures are looked into. New ideas relating to the LDPC decoding sum-product algorithm are then introduced and explored in a non-specific manner later in the thesis.

 

Status: Completed