Research on Adaptive Unscented Kalman Filter for Integrated Navigation
Tian Lai Xu · Advanced materials research · 2013
The accuracy of filtering deteriorates in condition that a priori information used in unscented Kalman filter (UKF) does not accord with the actual conditions. To improve the accuracy of filtering when the noise statistical properties are not known exactly in navigational data fusion, an adaptive UKF is proposed. In the filtering process, the statistical parameters of unknown system noises are adjusted online if filtering abnormality exists. Simulation results show that the proposed algorithm increases the accuracy compared with the standard UKF algorithm for integrated navigation.