Modeling and filter algorithm analysis of all-optical atomic spin gyroscope´s random drift

ARMA model is set up by using AIC method for all-optical atomic spin gyroscope´s random drift. The model is proven effective by the residual errors analysis. On the basis of the ARMA model, the results of conventional kalman filter, R-correction kalman filter, r-correction kalman filter and R-r correction kalman filter for analyzing gyroscope´s random drift have been compared in this paper. The results of actual datum analysis showed that R-correction kalman filter algorithm and R-r correction kalman filter algorithm performed better than others under high dynamic environment, conventional kalman filter algorithm performs better than others under low dynamic environment.