Inertial navigation system gyroscopic on-line calibration assisted by marine star sensor based on forgetting factor selection of a Sage-Husa filter

Abstract

In the process of ship autonomous navigation, the complex changes of the external environment lead to unknown statistical characteristics of star sensor measurement noise, which brings the estimation results of standard Kalman filtering divergence, and gyroscope on-line calibration cannot be effectively carried out. In view of the above problems, this paper proposed a gyroscope on-line calibration method assisted by star sensor based on forgetting factor selection of a Sage-Husa filter. According to the angular velocity, the output of gyroscope, and quaternion obtained from star sensor, the system error model with unknown noise statistical characteristics is established. Based on the judgment method of fuzzy logic, the error quaternion is used as the input, and the fuzzy rule is used to select the parameters of the adaptive filtering. Therefore, the measurement noise parameters of the filter are constantly estimated and corrected. Finally, the algorithm is verified by simulation test. The simulation results show that the proposed method can effectively estimate the gyroscope drift, scale factor under the case of changes in the external environment, and improve the accuracy and applicability of the navigation system.