The In-Operation Bias Drift Self-Calibration Method Of MEMS Gyroscope

Microelectromechanical(MEMS)gyroscope is an inertial sensor,which is applied to measure angular velocity or attitude angle.MEMS gyroscope has the advantages of small size,low cost,light weight and low power consumption.It is widely used in the civil automobile industry,industrial control,consumer electronics,military UAV,missile,fire control and other fields.The development aims of weapon systemsare miniaturization,portability and intelligence,which requires MEMS gyroscope to achieve high precision and high stability of measurement withsmall volume and low power consumption.However,due to the limitation of manufacturing process imperfection,material dispersion,air impermeability of package and other problems,it is a great technical and cost chanllenge tofurther improve the measurement accuracy and stability of MEMS gyroscope.In recent years,the digital control technology of MEMS gyroscopes was developed rapidly,and it has the characteristics of high stability,high precision and reconfiguration,which shows the great technical development advantages.By the application of intelligent self-calibration algorithm,thedigital control system of MEMS gyroscopes can break through the constraints of the state of the art to achieve high precision and high stability.Therefore,the research of intelligent self-calibration algorithm is important to expand the application fields of MEMS gyroscopes.In this dissertation,the bias error model based on the imperfections of processing technology and interface circuit of MEMS gyroscopes was established,and the bias drift self-calibration method based on the mode reversal and neural networkis proposed.To breaking through key technology of the self-calibrationcontrol system,the design methodology of digital control system for MEMS gyroscopes based on multi-objective parameter optimization is proposed.In order to verify the proposed theory model and algorithms,the hardware platform of bias drift self-calibration system for MEMS gyroscope is implemented.The experiment results show that the bias stability of MEMS gyroscopes has been effectively improved.The main contents of this dissertation are shown as follows:According to the bias error model of the stiffness imbalance,damping imbalance,mass imbalance,electrostatic excitation force imbalance,parasitic capacitive coupling and phase errorwhich are caused by the process imperfections and interface circuit errors,influence of all the above imperfections on the output signal of the fully symmetric MEMS gyroscope manufactured by our laboratory is analyzed in this dissertation,andthe feature of bias error signal of the MEMS gyroscope is also obtained,which provides the theoretical basis for the follow-up research on the bias drift self-calibration method and the parameter design of the digital control system.To calibrate the bias error of MEMS gyroscope,the bias drift self-calibration method based on mode reversal and neural network learning is proposed.More precisely,thereal-time mode reversal self-calibration method is applied to realize the preliminary calibration of the bias error signal.By applying the signal feature extraction and the threshold de-noising algorithm based on the improved CEEMDAN,the random noise in the bias drift signal after mode reversal is eliminated.Moreover,the neural network algorithm based on Bagging-ELM is adopted to compensate the residual erro ofbias driftsignal.The proposed self-calibration method significantly improves the bias stability of MEMS gyroscope.The design methodology of digital control system for MEMS gyroscope based on multi-objective parameter optimization is also presented in this dissertation.Through the multi-loop optimization method based on genetic algorithm and monte carlo analysis,the time comsumption of the design procedure of the closed-loop control loop is reduced,and the robustness of the control loop to the variations of the gyroscope structure parameters is improved.The least mean square demodulation algorithm based on adaptive moment estimation(Adam-LMSD)is proposed to improve the signal to noise ratio(SNR)of sense mode loop.Finally,the hardware platform of bias drift self-calibration system is implemented.The test results show that the bias instability of the tesed MEMS gyroscopes decreased from 23.76°/h and 19.8°/h to 2.7072°/h,the angle random walk decreased from 0.0071°/(?)and 0.0122°/(?)to 0.0013°/(?),the scale factor increased from 17.7 mV/(°/s)to 23 mV/(°/s),the scale factor's nonlinearity decreased from 84.52ppm to 61.56ppm,and the bandwidth increased from 23Hz to 101Hz.The novel and effective self-calibration methods for improving the bias stability of MEMS gyroscope are provided in this dissertation,which have some reference value for the follow-up research of high-precision MEMS gyroscope.