Online parameter identification and dynamic model reconstruction for AUV based on adaptive extended Kalman filter and...

Area of Science:

• Robotics
• Ocean Engineering
• Control Systems

Background:

• Accurate parameter identification is crucial for autonomous underwater vehicle (AUV) control.
• Ocean currents and measurement noise pose significant challenges to AUV parameter identification.
• Existing methods often struggle with complex dynamic environments and computational demands.

Purpose of the Study:

• To develop an integrated closed-loop framework for robust online parameter identification of a six-degree-of-freedom AUV.
• To enhance the accuracy and reliability of AUV state estimation and parameter identification under disturbances.
• To address computational burdens and improve robustness in dynamic underwater environments.

Main Methods:

• An advanced Adaptive Extended Kalman Filter with Prediction Error Method (AEKFPEM) algorithm was formulated.
• A fault-tolerant mechanism was embedded into the Extended Kalman Filter (EKF) with Sage-Husa noise estimation for enhanced robustness.
• A recursive update expression using the Sherman-Morrison-Woodbury formula was derived to reduce computational load.

Main Results:

• The AEKFPEM algorithm demonstrated robust online parameter identification capabilities.
• The integrated framework improved real-time state estimation and parameter identification accuracy.
• The method showed superior predictive performance in 5-second motion prediction experiments, even with model mismatch.

Conclusions:

• The proposed AEKFPEM algorithm offers a robust solution for online parameter identification of AUVs in challenging ocean conditions.
• The integrated closed-loop framework enhances AUV control system performance and reliability.
• The algorithm's effectiveness is validated through dynamic model reconstruction and prediction experiments, highlighting its practical utility.