Research On The Key Technology Of Redundant MEMS-IMU Error Compensation

Micro Electro Mechanical System(MEMS)has many advantages such as low cost,small size,low power consumption and high applicability.It has been widely used in aerospace,mobile communication,national defense technology and other fields.In the traditional three-axis quadrature Strapdown Inertial Navigation System(SINS),because the system does not have fault detection capability,when any accelerometer or gyroscope fails,the system will continue to output the navigation information with errors,causing unpredictable consequences.In consideration of cost and volume,the device level redundancy technology is usually used to improve the device level redundancy technology.Due to the diversity of configuration structure,the error model of redundant MEMS-IMU is poor in generality,and there are random and deterministic errors in MEMS devices that are susceptible to environmental interference.Therefore,redundant MEMS-IMU error analysis and calibration research is of great significance to improve navigation accuracy.Taking inertial navigation as the background and aiming at improving the output accuracy of redundant MEMS-IMU,this paper conducts research on key technologies such as error compensation and noise suppression of redundant MEMS-IMU.The research contents include the following aspects:In order to solve the problem of noise interference in MEMS inertial devices during signal acquisition,the noise reduction method based on improved wavelet threshold is studied.Firstly,Allan variance method is used to analyze the random error characteristics of MEMS inertial devices.Secondly,in order to solve the traditional wavelet threshold denoising algorithm in the hard threshold function is shock phenomenon,at the same time avoid the soft threshold function reconstruction error problem,put forward improved threshold function,by introducing a variable parameter,and adjusted the threshold function,enable it to both hard and soft thresholding function advantage,get rid of the noise in the protection of the signal at the same time can effectively avoid the traditional thresholding function caused by signal drift;Finally,an experiment of MEMS-IMU is designed to verify the effectiveness of the algorithm.Aiming at the problem of deterministic error compensation of redundant MEMS gyroscope,the research on redundant MEMS laboratory calibration scheme based on turntable is carried out.Firstly,the general error model of the redundant MEMS gyroscope is established by using the small-angle rotation vector method.Secondly,based on the traditional Kalman filtering calibration method,Extended Kalman Filter(EKF)suitable for nonlinear systems was adopted to improve the calibration accuracy in view of the non-linear error in the filtering results.Finally,aiming at the problem of slow convergence speed in the filtering process,an improved algorithm based on sequential filtering theory is proposed to make effective information enter the filtering earlier by changing the order of measurement information in the filtering process.The calibration accuracy and convergence speed of the algorithm are improved respectively from two aspects of nonlinear filtering and measurement information.Aiming at the problem that there is no rate reference provided by turntable in the field test of MEMS-IMU with low and medium precision,a self-calibration scheme of MEMS-IMU without the assistance of external equipment is proposed.The exploratory research on the calibration algorithm of accelerometer under the gravity vector at 9 positions was carried out,and a fast iterative algorithm based on the constant error deviation was proposed,which could realize the fast and accurate solution of the calibration parameters after two or three iterations,and the algorithm did not depend on the initial parameter setting.