An improved adaptive square root unscented Kalman filter for denoising IFOG signal

Abstract

An interferometric fiber optic Gyroscope (IFOG) is a core component in the inertial navigation system (INS), and used to measure the rotation rate of an object based on Sagnac principle. The output of IFOG suffers with noise and random drift errors, due to the variation and fluctuations of the ambient temperature during the operation time. Random drift error is the main source of error and it degrades the IFOG accuracy. To improve the precision of IFOG, the stochastic drift error models and noise compensation methods are required to suppress these errors. In this paper, the residual based an adaptive square root unscented Kalman filter (ASRUKF) is developed for denoising the IFOG signal. In this algorithm, the Kalman gain is adapted by using window average method and followed by covariance matching technique based on residual sequence. The proposed algorithm is utilized for IFOG test signal under static and dynamic environment. Allan variance (AV) analysis used to analyze and quantify the noise sources of IFOG sensor. In static and maneuvering condition, the performance improvement of proposed algorithm is indicated by the minimum values of variance and root mean square error (RMSE). A simulation result reveals that the proposed algorithm is a valid solution for drift denoising the IFOG signal as compared to Unscented Kalman filter (UKF).