| Autor: |
Romis de Faissol Attux, Murilo Loiola, Ricardo Suyama, Leandro de Castro, Fernando Von Zuben, João Romano |
| Zdroj: |
EURASIP Journal on Applied Signal Processing; Jul2003, Vol. 2003 Issue 8, p740-747, 8p |
| Abstrakt: |
This work proposes a framework to determine the optimalWiener equalizer by using an artificial immune network modeltogether with the constant modulus (CM) cost function. This studywas primarily motivated by recent theoretical results concerningthe CM criterion and its relation to the Wiener approach. Theproposed immune-based technique was tested under differentchannel models and filter orders, and benchmarked against aprocedure using a genetic algorithm with niching. The resultsdemonstrated that the proposed strategy has a clear superioritywhen compared with the more traditional technique. The proposedalgorithm presents interesting features from the perspective ofmultimodal search, being capable of determining the optimalWiener equalizer in most runs for all tested channels. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
