Quasi-optimal Algorithm for Receiving Discrete Signals based on Polygaussian Models

Abstract

The analysis of the use of quasi-optimal poly-Gaussian algorithms in the reception of discrete signals is carried out. It is shown that the noise immunity of quasi-optimal algorithms designed to combat chaotic impulse noise increases with the use of multi-account algorithms. The authors have developed a block diagram of a multi-threshold receiver for detecting discrete signals. It is noted that in the formation of the Polygauss model of the real signal-noise formation a priori leave a finite number of components, among which the algorithm in the process of work chooses smaller quantities, and at the last stage in the simplest case leaves only one hypothesis. In relation to Polygauss algorithms (in contrast to classical Bayesian hypotheses), we come to orthogonal models of input oscillation and to adaptation according to the parameters of the most plausible hypothesis. At the same time, along with this, simpler, technically convenient and quite effective algorithms with a priori rigid restriction of the set of hypotheses analyzed are possible. In addition, for example, algorithms with an intermediate degree of simplification are used. The work of a quasi-optimal gated pulse signal receiver against the background of noise and pulse noise is examined and analyzed. As a result of mathematical modeling, the dependences of the win providing a three-way solving scheme in comparison with a single-way one are obtained.