ARMA model based adaptive unscented fading Kalman filter for reducing drift of fiber optic gyroscope

Abstract

In this paper, the fiber optic gyroscope drift is modeled using an auto-regressive-moving-average (ARMA), time series model. The drift is subsequently reduced using the proposed adaptive unscented fading Kalman filter algorithm. The proposed algorithm has two cascaded stages for updating the state error and measurement noise covariance. In the first stage, the predicted state error covariance is updated using a transitive factor and in the second stage the measurement noise covariance is updated using another transitive factor. The suggested algorithm is used for reducing the drift of the FOG signal in both static and dynamic conditions at room temperature. The performance of the proposed algorithm is analysed using Allan Variance and drift for the static signal and root mean square error for the dynamic signal. The performance of the suggested algorithm is compared with the unscented Kalman filter (UKF) and a single transitive factor based adaptive UKF algorithm. The experimental results demonstrate that the proposed algorithm performs better than UKF and a single transitive factor based adaptive UKF algorithm for reducing the drift and random noise in both static as well as dynamic conditions.