An Improved Kalman Filter Based on Empirical Mode Decomposition for INS/CNS Navigation

This paper investigates the challenge of filtering for INS/CNS integrated navigation in the presence of uncertain measurement noise. To address this issue, the paper proposes an improved Kalman filter (IKF) that utilizes empirical mode decomposition (EMD). In the IKF, the high-frequency measurement noise is first obtained by the EMD of the measurement sequence. Then, the covariance of measurement noise is estimated using the high-frequency measurement noise to perform measurement updates for KF. Simulation results demonstrate that the proposed algorithm exhibits strong resilience against uncertain measurement noise and significantly improves the attitude accuracy of the INS/CNS integrated navigation system.