A performance degradation prediction method based on GPR and physics of degradation for MEMS gyroscope

MEMS (Micro-Electro-Mechanical System) Gyroscopes can monitor the attitude of aircraft to achieve stable flight, which is one of the most critical part of automatic flight. Performance estimation and prediction with high accuracy for the MEMS gyroscope can improve the stability and reliability of quadrotor UAV. In practical application, the degradation of material property, operating environment, etc will affect the performance of gyroscope, physics of degradation analysis can help understand the performance degradation process more clearly. Therefore, in this paper, a performance degradation prediction method based on Gaussian process regression (GPR) and physics of degradation modelling is developed. The mail performance indicators and degradation mechanism of MEMS gyroscopes are analysed, on this basis, the performance degradation process is simulated by Simulink. The prediction results based on one-dimension degradation data, two-dimension degradation data and three-dimension degradation data are compared, which proves that three-dimension degradation data can predict the remaining life of the MEMS gyroscope with high precision.