Probability of error in CPSK systems with intersymbol interference and additive noise

A simple lower bound, an upper bound and a simple approximation to the upper bound on the probability of error for coherent phase-shift-keyed (CPSK) systems operating in the presence of intersymbol interference and additive noise are obtained. The additive noises in the in-phase channel and the quadrature channel are assumed to be independent, and are independent of the signal, but not restricted to be Gaussian. The approximation to the upper bound is four times the lower bound, hence the tightness of these bounds is uniform for all cases. This fact and the simplicity of the bounds make these bounds a useful system design tool. Numerical examples for quaternary and octonary systems are presented and compared to known results.     

Probability of error in digital fiber optic systems with inter-symbol interference and signal dependent additive noise

Data transmission via optical fiber is a new discipline of communication theory. The principal difference from conventional baseband data transmission, which is characterized by a signal independent additive Gaussian noise, is the existence of a signal dependent shot noise.This paper presents a technique for estimating the error probability performance of digital systems with inter-symbol interference and signal dependent additive noise. For binary antipodal (±1) systems, the approximate upper bound to the error probability is twice the lower bound. Hence either can be taken as a good approximation to the actual error probability. The technique is then applied to a model of some promising optical data communication systems and a good approximation to the error probability is obtained. Some observations about the effect of various system parameters on the error probability and some numerical examples are presented.     

A class of generalized decision- feedback equalizers

An upper bound and a lower bound to the probability of error for data transmission systems using a decision-feedback equalizer, with error propagation effects taken into account, are presented. Numerical examples show that they are tight bounds.We introduce the concept of a generalized decision-feedback equalizer suggested by the form of the derived bounds. A subclass is studied and optimized numerically with the aid of the derived error probability bounds. The solution represents the best compromise achievable by that class between the output noise power and the output intersymbol interference due to error propagation. A numerical example is carried out in detail to illustrate the concept.