Federated Kalman Filter Based on KMEWMA Control Chart for Integrated Navigation

Integrated navigation systems are prone to faults in complex environments, necessitating the isolation or mitigation of their effects on the system. This paper introduces a federated Kalman filter (FKF) structure based on the kernel multivariate exponentially weighted moving-average (KMEWMA) control chart, which incorporates statistical process control techniques to detect and mitigate faults effectively in integrated navigation systems. The proposed method demonstrates outstanding fault detection capability, successfully addressing both gradual and sudden-varying faults. Compared with chi-square test, the KMEWMA control chart could detect faults more precisely. Furthermore, instead of solely isolating faults, the system adaptively adjusts them to enhance positioning accuracy and stability. The simulation results validate the effectiveness and superiority of the proposed method.