Research MEMS Sensor Calibration And Compensation Algorithms In IMU

The cost of MEMS inertial components' production is low, which has been widely used in various fields for the quality and volume are more portable. The precision of the micro inertial measurement unit is still insufficient due to the error exist in the MEMS inertial device. The effectiveness of both navigation system and navigation position system mainly depends on the precision of inertial measurement unit(IMU). The measurement precision of the system can be improved effectively by mathematics error modeling compensation and correcting various errors of the MEMS inertial device. Therefore, establish mathematical model for MEMS inertial device accurately is an important part in the field of inertial attitude measuring.Firstly, development of micro inertial is summarized in this thesis, research significance and research situation of MEMS inertial device error at home and abroad are introduced. Then the working principle of MEMS gyroscope and accelerometer and key performance indicators are introduced respectively. The error source of the micro inertial measurement system is analyzed and the error formation mechanism of inertial device can be analyzed,such as deterministic error, random error and temperature error and so on.Secondly,temperature error compensation model is established, as well as temperature drift and hysteresis phenomenon of MEMS inertial device are analyzed. Corresponding temperature test is designed and the temperature characteristic of MEMS inertial devices is analyzed. Gyroscope is compensated using BP neural network model, which combined rotation speed and temperature. Y axis is compensated using the method of least square based on the temperature characteristics of accelerometer. Zero temperature error of the compensated gyroscope is improved one order of magnitude compared with traditional compensation method, and the zero bias error coefficient of accelerometer was 7.5 mg /°C before, and it rise to 3.2 e-2 mg/°C.Thirdly, the certainty error is modeled and calibrated. Certainty error of MEMS inertial components including zero offset error, scale factor error and setting error and so on is corrected by the improved traditional method. The average axis offset of the MEMS gyroscope compensated reduced to 0.04%, 0.03%, 0.01%, respectively, and the mean absolute error of pitching angle calculated through MEMS accelerometer was 1 ° before, and it increases to 0.14° after compensating.Finally, random noise error of MEMS inertial device is compensated, noise of gyroscope and accelerometer are analyzed by using the Allan variance, and model is set by utilizing time series method, Kalman filter is used for signal noise suppression and the filtering effect is well. After compensation, the zero bias instability of MEMS gyroscope and accelerometer increased by 65.7% and 52.8% respectively.