A MLP/BP-based equalizer for NRZ signal recovery in band-limited channels

A multi-layered perceptron neural network with the backpropagation algorithm (MLP/BP) is realized as an equalizer for nonreturn-to-zero (NRZ) signal recovery in band-limited channels. It is applied as an adaptive filter to recover the NRZ signal. By computer simulations, the proposed design can recover 100 MHz NRZ data in a 10 MHz channel.

[1]  A.M. Davis,et al.  Microelectronic circuits , 1983, Proceedings of the IEEE.

[2]  Benjamin Van Roy,et al.  Solving Data Mining Problems Through Pattern Recognition , 1997 .

[3]  Sang-Won Nam,et al.  Channel equalization for severe intersymbol interference and nonlinearity with a radial basis function neural network , 1999, IJCNN'99. International Joint Conference on Neural Networks. Proceedings (Cat. No.99CH36339).

[4]  A. W. M. van den Enden,et al.  Discrete Time Signal Processing , 1989 .

[5]  A. Sedra.  Microelectronic circuits , 1982 .

[6]  Bang-Sup Song,et al.  NRZ timing recovery technique for band limited channels , 1996, 1996 IEEE International Solid-State Circuits Conference. Digest of TEchnical Papers, ISSCC.