The purpose of this paper is to describe a typical communications system and highlight the problems encountered when transmitting signals. We shall discuss how the signal is corrupted during transmission and examine the possibility of a telecommunications system within which the transmitting and receiving terminals continually adapt their strategies to optimise the use of the particular communications channel at their disposal.

There are many ways in which a signal can be corrupted during its transmission from a sending terminal to a receiving terminal. Figure 1 shows a generalised system and indicates where and in what way the signal can be corrupted and, in particular the various forms of noise that can be added to it. Within this generalised setting, the original signal may first be sent to a switching centre. The switching centre will also have a variety of other signals from other terminals arriving simultaneously and a common form of noise/interface is caused by ‘crosstalk’ which involves the ‘leakage’ of one signal to another. The next processing step will probably involve modulation.

There are two main reasons for using modulation. These are firstly to shift the message signal frequencies to a band which can be transmitted efficiently over the channel and secondly to enable the simultaneous transmission of several signals over a single transmission link. In order to understand the significance of this process one must appreciate that if the frequency domain of a time–varying signal is examined one normally finds that most of the information is contained in a finite portion of the frequency spectrum.