Neural Network And Its Angular Rate Sensor Calibration Applications

Silicon gyroscope has low precision and serious non-linear errors. In order to improve the precision of the Silicon gyroscope, need a method to compensate it. Neural network has the ability of approaching nonlinear function at random precision, so that correct non-linear errors of angular rate sensor on the basis of neural network method.This paper analyses, designs research scheme of the subject at first, puts up a simple experimental flat, gathers datum with two kinds of precision needed to obtain sample data sets, sends into neural network system designed. Through studying and data analysis, designs better compensation model based on neural network. The paper tries BP and RBF two kinds of commonly used neural networks compensation model mainly, has compared with many kinds of networks and learning algorithms through the instance. Experimental results indicate, using neural network methods can realize the function of the filter, and improve the precision of the sensor.