Improved cubature Kalman filtering for tightly coupled GPS/SINS integrated navigation system

Abstract

In the integrated navigation system, unscented Kalman filter (UKF) has been widely used. Compared to traditional Kalman filter and extended Kalman filter, the UKF algorithm has improved in accuracy than Kalman filter. However, UKF has no strict mathematical proof, and when the order of the system increases, the error of UKF becomes larger, even cause divergence. In this paper higher order cubature Kalman filter is introduce in GPS/SINS tightly coupled navigation system to improve the accuracy and stability. And the accuracy of higher order cubature Kalman filter and UKF are analysed through Taylor expansion in this paper, which proved higher order cubature Kalman filter performs better than UKF in mathematics. The simulation results show that the improved cubature Kalman filter can effectively improve the navigation system's accuracy and stability.