Communications equalisation using adaptive techniques

The process of distortion in communications systems may, typically, be modelled as a linear process between the transmitter and the receiver. A common means of overcoming this problem has thus been to introduce a (linear) inverse filter into the receiver to `equalise´ the channel. However, the designer does suffer from the disadvantage of having no a-priori knowledge of the channel transfer function and, is further assured that this function will be time varying. For these reasons such equalisers are commonly adaptive in nature; that is, they automatically adjust their parameters subject to some external (statistical) stimuli. Common techniques which are applied to this problem deploy linear adaptive filters in an attempt to combat these problems. Such techniques are popular due to the highly developed nature of linear adaptive filter theory, and the relative simplicity of such processes. However, it will be shown here that this problem is one which is inherently non-linear in nature and an alternative strategy for equalisation is evolved