Maximum correntropy generalized high-degree cubature Kalman filter with application to the attitude determination system of missile

Abstract

For SINS/CNS integrated navigation system, the CKF can perform well for state estimation in Gaussian noise. However, its performances are likely to degrade significantly under non-Gaussian noise conditions. To improve the robustness of the CKF against non-Gaussian noise, we propose an improved cubature Kalman filter, called the maximum correntropy generalized high-degree CKF (MCGHCKF). In the MCGHCKF, the generalized high-degree cubature rule is used to improve the filtering performance, and the maximum correntropy criterion is utilized to reduce the influence of non-Gaussian noise on state estimation. Simulation experiments illustrate the effectiveness and robustness of our algorithm.