A new type of adaptive kalman filtering algorithm and its application

In order to avoid the divergence and improve the real-time property of the filter, an adaptive Kalman Filtering (AKF) algorithm is presented, which obtained a factor of filtering state by using the criterion of filtering anomalies, and the measurement noise covariance matrix is confirmed by using this factor. AFK is realized by changing the measurement noise covariance according to the state of the filtering. AKF is applied to the INS/DS (Inertial Navigation System/Double-star System) integrated navigation system, and compared it with the conventional Kalman filtering and simplified Sage-Husa filtering. The simulation results showed that the AKF simplified the computation and improved the real-time property with the same accuracy of simplified Sage-Husa filtering.