Nonlinear adaptive equalisation [multilayer perceptron]

Develops the concept of the linear equaliser as a classification process within the domain of digital communications equalisation. From this point it is demonstrated that nonlinear adaptive structures are in fact more suited to solving this classification problem. The result of applying this concept is illustrated by developing a very general nonlinear adaptive structure in the form of the multilayer perceptron. Simple examples are provided which clearly illustrate the advantageous performance compared to linear equivalents