The application of RBF neural network in the compensation for temperature drift of the silicon pressure sensor

Abstract

Temperature drift is the important factor of the precision of diffused silicon pressure sensor, so author uses software to compensate for it to improve the precision of the sensor. At the data base of the temperature characteristic experiment of diffused silicon pressure sensor, author proposes to use RBF neural network to establish temperature drift compensated model with regression analysis. Compared with two-dimension regression analysis, RBF neural network can improve the precision of the model distinctly.