Abstract
It is shown that, to measure the velocity of motion of a plasma in the channel of an electrodynamic accelerator, the use of artificial neural networks is promising, since they enable one to approximate the mathematical functions with minimum error when using the coordinate-function method. The use of a multilayer perceptron with activation functions of the neurons of the “rational sigmoid” type is proposed as a basis for constructing a neural network system for measuring velocity.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
REFERENCES
E. V. Kirievskii, Izv. Vyssh. Ucheb. Zaved. Elektromekh., No. 4, 74 (2000).
F. Wasserman, Neurocomputer Techniques: Theory and Practice [Russian translation], Mir, Moscow (1992).
P. M. Kolesnikov, Electrodynamic Acceleration of a Plasma [in Russian], Atomizdat, Moscow (1971).
D. E. Sinel'nikov and E. M. Sinel'nikov, Izv. Vyssh. Ucheb. Zaved. Elektromekh., No. 7, 10, (1983).
V. E. Kirievskii and E. V. Kirievskii, Certificate No. 2000611373 on the Official Registration of a Computer Program. Recorded in the Register of Computer Programs, December 25, 2000.
V. G. Tsaregorodtsev, Proceedings of the Sixth All-Russia Seminar “Neuroinformatics and Its Applications,” Krasnoyarsk State Technical University, Krasnoyarsk (1998), p. 207.
Rights and permissions
About this article
Cite this article
Kirievskii, V.E., Kirievskii, E.V. An Analysis of Neural Network Structures of a System for Measuring the Rate of Acceleration of Bodies in an Electrodynamic Accelerator. Measurement Techniques 47, 58–65 (2004). https://doi.org/10.1023/B:METE.0000022505.33978.38
Issue Date:
DOI: https://doi.org/10.1023/B:METE.0000022505.33978.38