Комп’ютерне моделювання поліноміальних двофункціональних правил обробки радіосигналів на фоні негаусівських завад

Abstract

Проведено комп’ютерне моделювання поліноміальних двофункціональних правил обробки радіосигналів на фоні асиметрично-ексцесної завади другого типу першого виду.

In this paper using two functional rules for processing of incoming sample values (1) is proposed. The first function is the function of hypothesis distinction, based on polynomial decisive rules. The optimal coefficients of such decisive rules are obtained by moments criterion of the upper limit of probability of errors. The second function is the function of estimation of signals parameters, based on the method of polynomial maximization and method of truncated polynomial maximization. The two functional rules for signal processing use the moment and cumulant description of random variables. Using the traditional approach to research and develop systems of processing of Non-Gaussian random processes is characterized by significant limitations associated with the complexity of algorithmic implementation and increasing computational resources that leads to appropriate difficulties in establishing quality software-algorithm and hardware means of signal processing. In this regard, there is an urgent task to develop effective methods of signal processing to enhance the accuracy of Non-Gaussian signals processing compared with the traditional correlation approach. Computer modeling of polynomial two functional rules to process radio signals at skewness-kurtosis non-gaussian interference is conducted. The skewness-kurtosis interference is characterized by cumulant coefficients of the third and fourth order. Experimental results of computer modeling of distinction and estimation algorithms in general conform to theoretical. It is shown (Fig. 1-3) that the account of a thin structure of NonGaussian interferences can increase efficiency of signal processing algorithms. The greatest efficiency is reached for boundary conditions of the coefficients of skewness and kurtosis of the interference and for higher degrees of the polynomial. The obtained results may be used to reduce the probabilities of errors of distinction of radio signals and improve the accuracy of the estimations of their parameters in radiolocation, radio navigation and other areas