DETECTING OF THE DETERMINISTIC SIGNALS ON A PHON OF GAUSSIAN NOISE WITH USE OF NEURAL NETWORK LOGIC BASIS

V. Ryzhakov, Surgut state university, Surgut, Russia

SYNCHROINFO JOURNAL. Volume 5, Number 3 (2019). P. 18-21.

Abstract

In this paper the neural network solving a problem of detection of the deterministic signal on a phon of noise in conditions of parametrical aprioristic uncertainty is described. The neural network is constructed proceeding from the assumption that the noise signal submits to the normal law of distribution. Besides the algorithm of training of the constructed network that is modification of error back propagation algorithm is offered.

Keywords: attitude of credibility, density of probability, normal law of distribution, deciding rule, threshold of decision-making, neuron, neural network, training sample, gradient method of optimization, algorithm of neural network training.

References

[1] The theory of detection of signals. P.C. Akimov and other; General edition is P.A. Bakut. Publishing house “Radio and communication”, Moscow, 1984.
[2] Levin B.R. Theoretical bases of a statistical radio engineering. Publishing house “Radio and communication”, Moscow, 1989.
[3] A.I. Galushkin. The theory of neural networks. Vol. 1.: The manual for high schools. The publishing enterprise of edition of magazine “Radio engineering”, Moscow, 2000.
[4] Neural mathematics. Vol. 6. The manual for high schools. Ageev A.D. and other; General edition is A.I. Galushkin. The publishing enterprise of edition of magazine “radio engineering”, Moscow, 2002.