Slow-Varying Measurement Outlier Mitigation Method for HVC-Assisted SINS/DVL Navigation Applications

Underwater Doppler Velocity Log (DVL) measurement is susceptible to slow changes owing to interference from forced maneuvers, ditches and fish shoals under uncharted ocean environments, where mismatched measurement model may affect navigation performance of SINS/DVL navigation applications. To solve such problems, an enhanced Kalman filtering method based on adaptive measurement noise estimation technique is proposed in this paper. By designing a centripetal acceleration errors-reinforced holonomic velocity constraint (HVC) and the adaptive Sage-Husa approach, the parameters of the Kalman filter are adjusted online to maintain the system stability and filtering accuracy, which contributes to the measurement outlier mitigation in SINS/DVL navigation applications. The sea trial of autonomous underwater glider (AUG) reveals the proposed method is resistant to external disturbances and maintains the filter convergence in uncharted ocean environments while providing more accurate position estimates than existing methods.