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.

No abstract

A blind identification algorithm f o r unknown linear channels is proposed. The algorithm operates ower a grid in the channel space that is made finer by using the Maximum Likelihood criterion lo confine the estimated channel in the neighborhood of the original unknown channel. The nature of the algorithm leads t o efficient parallel implementation, and its storage requirements are only those of the Viterbi algorithm. A s is demonstrated in the simulation examples, lengths of 50-100 samples are enough to identify channels when binary (PAM-2) transmission is used. When multilevel signaling is used, a reduced constellation approach can provide convergence with data blocks of length 200-500 samples.

No abstract

No abstract

No abstract

No abstract

No abstract