Share Email Print

Proceedings Paper

Neural network technologies in experimental computer science (measurement, monitoring, diagnostics, and tests)
Author(s): Vladyslav D. Tsydelko; Sergiy O. Bragynets
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Distinguished tendency of modern experimental informatics (EI) and metrology technical means development is their intellectualization. Examples of neural network technologies (NNT) application in modern EI for measurement, monitoring, diagnostics and tests are indicated. Features of NNT from the point of view of their application in means of EI are considered, field of solved by them problems is determined. Parallel 'measurement - classification' is conducted, i.e. problem of measurement -- determination of numerical value of physical quantity -- is considered as problem of classification. Possibility of NNT application in EI for interpolation of interferometer quadrature measuring signals with the purpose of resolving possibility increase in range of nanometer is considered. For solution of interpolation problem some structures of artificial neural networks such as perceptron and probabilistic networks are analyzed. Comparative analysis of exactitude for all considered structures is indicated. Increasing of the interferometer resolution by means of NNT application is possible theoretically up to 0.001 nm.

Paper Details

Date Published: 12 June 2001
PDF: 6 pages
Proc. SPIE 4425, Selected Papers from the International Conference on Optoelectronic Information Technologies, (12 June 2001); doi: 10.1117/12.429716
Show Author Affiliations
Vladyslav D. Tsydelko, National Technical Univ. of Ukraine (Ukraine)
Sergiy O. Bragynets, National Technical Univ. of Ukraine (Ukraine)

Published in SPIE Proceedings Vol. 4425:
Selected Papers from the International Conference on Optoelectronic Information Technologies
Sergey V. Svechnikov; Volodymyr P. Kojemiako; Sergey A. Kostyukevych, Editor(s)

© SPIE. Terms of Use
Back to Top