Using recurrent neural networks for adaptive communication channel equalization

Abstract

Nonlinear adaptive filters based on a variety of neural network models have been used successfully for system identification and noise-cancellation in a wide class of applications. An important problem in data communications is that of channel equalization, i.e., the removal of interferences introduced by linear or nonlinear message corrupting mechanisms, so that the originally transmitted symbols can be recovered correctly at the receiver. In this paper we introduce an adaptive recurrent neural network (RNN) based equalizer whose small size and high performance makes it suitable for high-speed channel equalization. We propose RNN based structures for both trained adaptation and blind equalization, and we evaluate their performance via extensive simulations for a variety of signal modulations and communication channel models. It is shown that the RNN equalizers have comparable performance with traditional linear filter based equalizers when the channel interferences are relatively mild, and that they outperform them by several orders of magnitude when either the channel's transfer function has spectral nulls or severe nonlinear distortion is present. In addition, the small-size RNN equalizers, being essentially generalized IIR filters, are shown to outperform multilayer perceptron equalizers of larger computational complexity in linear and nonlinear channel equalization cases.

Uncertainty, fuzzy logic, and signal processing

“…Simulations were run for 5 di!erent SNRs ranging from 26 to 34dB (26:2:34). Note that we ran our simulations for high SNRs, because the channel in (26) is very nonlinear, and because the simulations in [21] were presented for SNR ranging from 22 to 40 dB. We ran 50 Monte Carlo simulations for each value of SNR.…”

Complex discriminative learning Bayesian neural equalizer

“…Performance of the DL equalizer (DLE) was tested for di erent channel models [4,9,10]. Results obtained by an equalizer having the same DF architecture, but trained with the traditional MSE criterion, were also considered for comparison.…”

“…with being a positive number and D i (y) is given by (10). Formula (13) assumes positive values in correspondence of each wrong decision, and negative values in the opposite case.…”

“…Alternatively, multilayer feedforward [1,5,9,14] or recurrent [10,12] neural networks can be used as classiÿers, using the mean square error (MSE) criterion. However, MSE minimization does not imply a reduced bit error rate (BER) [9] and alternative error functionals may be devised.…”

Optimized Decision Feedback Equalizer Algorithm based on Sparse Underwater Acoustic Channel

“…Nonlinear adaptive filters based on a variety of neural network models was used successfully for system identification and noise-cancellation in a wide class of applications. An adaptive Recurrent Neural Network (RNN) based equalizer whose small size and high performance makes it suitable for high-speed communication channel equalization (Kechriotis, Zervas and Manolakos, 1994;Fang and Chow, 1999). Some novel blind equalization approach based on radial basis function (RBF) neural networks are proposed.…”