View article

The emulation of a decision feedback equalizer (DFE), operating on a noiseless finite impulse response channel, by a feedforward multilayer processor (similar to a neural network) is considered. The similarity between the two systems is exploited to obtain tight bounds on the probability of error, as a function of the number of layers, using the theory of finite state Markov processes. A class of channels for which exact representation by a neural network of finite complexity is possible is established. Two training algorithms for the adaptive DFE emulator, the first for sigmoid and the second for hard limiting processing elements, are derived and evaluated by simulation.