Research on modeling and filtering method of atomic spin gyroscope's random drift

Abstract

Atomic spin gyroscope is a new kind of gyro based on quantum mechanics, with ultra-high precision, simple structure, small size, etc. Therefore, to study the characteristics of random drift for improving the accuracy of atomic spin gyro is significant. Firstly, through the analysis of the gyro static output data and preprocessing, the stationary time series gyro random error has obtained. Then established the gyro drift error model based on time series ARIMA, and test the applicability of the model through analysis of residual. Finally, through the establishment of the Kalman filter based on the model to filter out the gyro random drift, and the use of Allan variance analysis of the filtering results. The results show that the modeling and filtering method and can effectively reduce the atomic spin gyro random drift error, the angle random walk coefficient raised an order of magnitude compared with traditional ARMA modeling method, so as to effectively enhance the output accuracy of atomic spin gyro and stability of the system.