Methods and models of neural networks for approximation of calibration characteristics of NTC-thermistors

Autor:	Serhii Fedin, Irina Zubretska
Rok vydání:	2022
Předmět:	Computational Theory and Mathematics Artificial Intelligence Applied Mathematics Theoretical Computer Science
Zdroj:	System research and information technologies. :102-120
ISSN:	2308-8893 1681-6048
DOI:	10.20535/srit.2308-8893.2022.3.07
Popis:	The hypothesis about the expediency of using RBF-networks to improve the accuracy of constructing the calibration characteristics of NTC-thermistors in the operating temperature range without dividing it into subranges is confirmed. It has been established that the error of the neural network approximation of the calibration characteristics of NTC-thermistors based on RBF-networks is at least one and a half times less than the permissible error of approximation of the third-order polynomial model, which is used in the software of modern systems for collecting and processing measurement information. A technique has been developed for processing measurement information using adaptive RBF-networks to automate constructing individual calibration characteristics and periodic calibration of NTC-thermistors.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::216732eed32ec0b543fcf50681617766 https://doi.org/10.20535/srit.2308-8893.2022.3.07 Zobrazit plný text záznamu