Modeling and Sectional Compensation of Temperature Error of FOG Based on BOA-GBDT
Cao Feng, Yan Ting-Yang, Su Min · 2021
According to the requirements of temperature compensation for real-time and accuracy, a method of using Bayesian algorithm to optimize the gradient boosting tree regression is proposed to establish the temperature error compensation model of fiber optic gyroscope, and it adopts the method of real-time acquisition of temperature change rate with multiple data windows to meet the requirements of online compensation and model input. The fiber optic gyroscope is placed in a temperature box to perform a temperature change test of −40-60°C to obtain measured data. The temperature and temperature change rate are used as input, and Bayesian algorithm optimization gradient lifting tree regression modeling and temperature rising and falling segment modeling are performed respectively. The comparative experiment results show that the proposed model achieves the best compensation effect. Through the compensation comparison test, it is verified that the proposed model has good compensation ability and generalization ability for non-training data.