Soft Threshold Test Based Dual-Loop Control for Underwater Vehicle in Complex Environments

This paper introduces the challenge of accurately acquiring self-state information required for underwater position-velocity dual-loop control of Unmanned Underwater Vehicle (UUV) in complex underwater environments. A method based on a soft threshold chi-square detection adaptive Kalman filter is proposed to perform filtering processing on the full-dimensional measurements of UUVs. This approach effectively corrects velocity and position state variables by incorporating kinematic correlations and constructing weighting metrics to achieve optimal utilization of suspicious measurements. Simulation results demonstrate that integrating the filtered information into the dual-loop control framework significantly smooths state estimation outcomes and mitigates oscillation risks induced by control actions.