Hybrid Extended Kalman Filtering Triggered by the Degree of Polarization for INS/BPNS Integration

To handle the non-Gaussian measurement noise involved in the INS/BPNS (inertial navigation system/bionic polarization navigation system) integration, this paper proposes a hybrid extend Kalman filtering (EKF). Firstly, a new mode of the INS/BPNS integration is constructed by using the relationship between the E-vector angle and INS misalignment angle based on the inherent property of the polarized E-vector perpendicular to the scattering plane; Further, a maximum correntropy criterion-based EKF (MCC-EKF) is established based on the MCC theory to address the influence of non-Gaussian measurement noise; Based on above, a hybrid EKF is established, which is triggered between the EKF and MCC-EKF by the BPNS measurement’s degree of polarization to reduce the unnecessary computational burden. Simulation of INS/BPNS integration for an unmanned surface vessel navigation shows the proposed hybrid EKF can improve the navigation accuracy by curbing the adverse effects of the non-Gaussian measurement noise compared with the traditional EKF.