A Markov Model for nonlinear channels with memory and some applications

A Markov chain model is proposed for the analysis of bandlimited, and possibly nonlinear, digital communication channels. The fundamental feature of this approach is its analytical tractability, as well as its flexibility. From this model several results can be obtained in a relatively straightforward way. An application considered is the evaluation of the signal power spectrum at the channel output. The bandwidth occupancy, and hence the interference among adjacent channels, in a multiple-user communication situation subject to spectrum spread due to a nonlinearity (a typical example occurs in digital satellite communication systems) is evaluated. As another application, the structure of the optimum (maximum likelihood sequence) receiver for a channel modeled in this way is derived, and a Viterbi algorithm is set up for the demodulation of a digital signal.