Analysis and compensation of MEMS gyroscope drift

Compared with traditional rate sensors, the Micro Electro Mechanical System (MEMS) gyroscope is smaller, lighter, cheaper and lower in power consumption. Thus, it has been widely used in consumer electronics, automation electronics and inertial navigation systems. However, because of the limitations of contemporary technology, the MEMS gyroscope usually has structure defect which will result in large drift. In this paper, we will report a method for analysis and compensation of MEMS gyroscope drift. Firstly, the gyro error characters are analyzed with the help of Allan variance. After that, Autoregressive (AR) model and auto-regressive moving-average (ARMA) model of gyroscope random drift are built. Besides, Kalman filter will be designed to decrease the random error of gyro. Test result shows that the variance of KF output is no more than 30% of the original signal variance. In conclusion, the combination of Allan variance, time series analysis and Kalman filter induce an excellent compensation effect of MEMS gyroscope drift.