In this paper we introduce an enhanced Decision Feedback Equalizer (DFE), based on the use of a feedforward neural network trained with the Discriminative Least Squares (DLS) algorithm. The DFE is a very common architecture in communications [4] ; its ability to cope with channels characterized by a high Intersymbol Interference (ISI) comes from the degree of nonlinearity and the feeback introduced. In this work we show how Neural Networks can generalize the DFE, giving superior performance in the presence of non-minimum phase and non-linear channels. In this last case, the Neural DFE (NDFE) outperforms a Viterbi decoder with a decision depth of five symbols.
Improved decision feedback equalizer using discriminative neural learning / Cocchi, F. F.; DI CLAUDIO, Elio; Parisi, Raffaele; Orlandi, Gianni. - STAMPA. - (1998), pp. 623-625. (Intervento presentato al convegno IEEE 1998 International Conference on Neural Networks and Brain ICNN&B'98 tenutosi a Beijing, China nel October 27-30, 1998).
Improved decision feedback equalizer using discriminative neural learning
DI CLAUDIO, Elio;PARISI, Raffaele;ORLANDI, Gianni
1998
Abstract
In this paper we introduce an enhanced Decision Feedback Equalizer (DFE), based on the use of a feedforward neural network trained with the Discriminative Least Squares (DLS) algorithm. The DFE is a very common architecture in communications [4] ; its ability to cope with channels characterized by a high Intersymbol Interference (ISI) comes from the degree of nonlinearity and the feeback introduced. In this work we show how Neural Networks can generalize the DFE, giving superior performance in the presence of non-minimum phase and non-linear channels. In this last case, the Neural DFE (NDFE) outperforms a Viterbi decoder with a decision depth of five symbols.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.