Generalized Maximum Correntropy Cubature Kalman Filter with Variational Bayesian for SINS/GPS Integrated Navigation Systems

To address the degraded accuracy and poor robustness of Strapdown Inertial Navigation Systems (SINSs)/Global Positioning Systems (GPSs) integrated navigation systems under time-varying non-Gaussian measurement noises, this paper proposes a variational Bayesian generalized maximum correntropy cubature Kalman filter (VBGMCCKF). The proposed method combines variational Bayesian adaptive method with the generalized maximum correntropy criterion, enabling the filter to handle the noises with time-varying statistical characteristics and effectively improving its applicability to different types of non-Gaussian noises. The results under different scenarios demonstrate that VBGMCCKF achieves superior estimation accuracy and robustness in the SINS/GPS integrated navigation systems compared with other existing methods. These results confirm the effectiveness of the proposed method for integrated navigation systems under complex noise environments.