Modeling and compensation method for temperature error of laser gyroscope based on support vector machine

Laser gyroscope error determines the measurement accuracy of position and orientation system(POS)used for airborne SAR motion compensation. The problem of the severe influence of environmental temperature on laser gyroscope measurement accuracy is studied; the temperature error mathematical model of laser gyroscope is analyzed. Based on the function fitting capability of support vector machine(SVM), a new modeling and compensation method for gyroscope temperature error is proposed. Compared with the compensation methods based on linear regression or neural network, the new method has the advantages of higher precision and generalization performance. Experiment results demonstrate that the proposed method could decrease the temperature error of laser gyroscope by 75% and promote its adaptability of laser gyroscope towards environmental temperature.