Using interval neural network prediction for measuring instruments calibration in control systems

The possibility is analyzed of using neural network prediction results for calibrating the measuring instruments, being the part of control systems. The modification of quality functional is given and the use of the interval global optimization algorithm is suggested for interval neural network models training. The developed software is described in brief. Considered are the calibration technique and the measurement procedure of determining the interval neural network prediction applicability. The results of computing experiments are given.