Reserch On Stochastic Error Analysis,Modeling And Filtering Of Gyro

With the development of the manufacturing industry,the accuracy of the gyroscope is constantly getting improved,and the accuracy of the inertial navigation system is constantly increasing.With the improvement of the manufacturing precision of gyroscopes,the influence of random errors on the accuracy of gyroscopes must not be ignored.In order to reduce the influence of random error on navigation accuracy,the random errors of gyro are studied with MEMS gyro,laser gyro and fiber optic gyro as objects.Gyro random error analysis methods mainly include autocorrelation function method,power spectral density method and Allan variance analysis method.In order to analyze the random errors of the three kinds of gyroscopes,the subject first analyzes their autocorrelation functions and power spectral density,and then uses the Allan variance analysis method to analyze their random errors.However,in order to obtain the noise coefficients of three kinds of gyro random errors,the curve of Allan variance analysis results was fitted,and then the noise coefficients of gyro random error were obtained.Finally,in order to verify the influence of total sampling time,different times of power on,and cut-off frequency of the filter on the gyro random error Allan variance anal ysis,experiments were carried out,and then the experiment data was analyzed by Allan variance.Although Allan variance analysis method obtains the noise coefficient s of gyro random error,it has not been used to establish a random error model for gyro ou tput.In order to establish the ARMA model of gyro random error,the stability of the gyro random error signal is first tested,and then based on the test results,the recursive augmented least squares method is used to estimate the model parameters of the gyro random error ARMA model.The applicability of the established ARMA model was verified by AIC criteria.The ARMA model was rewritten into a discrete state space model,and the Kalman Filter technique was used to filter the gyro output.The output data of the gyro before and after filtering were compared,and the Allan variance analysis results of the data before and after filtering were also compared and the corresponding results were obtained.Finally,in order to facilitate the analysis,modeling and filtering of gyro random errors,a gyro output data acquisition platform was built.The experimental platform uses RS422 serial port protocol to communicate.It can flexibly set the format of data frames and complete the data collection and Allan variance analysis of the three gyro outputs of different protocols.