Detection of discontinuous signals from correlation processing of input realisations

Abstract

Despite the in-depth study of methods for detecting radio emissions under conditions of a priori uncertainty about their parameters, only asymptotic expressions for discontinuous signals have been obtained, the application of which is limited by various assumptions. An analysis of practical solutions for estimating the time interval characterizing the position of discontinuous signals in the processed sample showed that they are based on correlation processing methods that allow obtaining acceptable results under the assumption that the channel noise is approximated by a Gaussian distribution. Under such conditions, the fragments of the input realisations containing the detectable signal are characterised by a sufficiently high variance, which increases the efficiency of their subsequent processing. In conditions of high noise levels, however, this effect is not so pronounced. This leads to the need for additional use of Fourier transform procedures, which is not always acceptable in practice. To develop a method of detection of discontinuous signals on the basis of correlation processing of input realizations in conditions of a priori uncertainty about their parameters. We have developed a threshold detector device with two correlators connected in series and allowing us to increase the signal-to-noise ratio in the processed realizations. The expressions for calculating the probability estimates of false alarms and target miss when detecting discontinuous signals have been presented. The analysis of obtained results has shown that for discontinuous signals the significant problem is a high value of false alarm pro-bability. The results have shown the effectiveness of the proposed approach to the detection of discontinuous signals based on detectors that implement serial calculation of the cross-correlation functions. The simulation results confirm the conclusion that it is necessary to increase the variance in the processed sample through the use of multiple correlators. Thus, the additional correlation processing allows to increase efficiency of algorithms of detection of discontinuous signals.