Random Error Compensation of MEMS Gyroscope Based on Adaptive Kalman Filter

Aiming at the low precision and large random drift of low-cost MEMS gyroscopes, an adaptive Kalman filter compensation method based on Singer model is proposed. Different from the commonly used time series modeling methods, the Singer model is introduced into the random drift modeling of MEMS gyro, which avoids the tedious steps of time series modeling. Based on the model, the adaptive Kalman filter based on Allan variance is designed. The simulation results show that the Singer model and the adaptive Kalman filter based on Allan variance can effectively reduce the random error of MEMS gyroscopes, and verify the feasibility of the algorithm and have certain reference value.