Adaptive Robust EKF with NARX-Based Velocity Prediction for High Precision AUV Navigation Under DVL Outages

Area of Science:

• Robotics
• Oceanography
• Navigation Systems

Background:

• Autonomous Underwater Vehicles (AUVs) face navigation challenges in complex underwater environments due to sensor noise and outages.
• Doppler Velocity Log (DVL) systems, crucial for AUV navigation, are prone to outages, leading to degraded performance.

Purpose of the Study:

• To develop an integrated navigation framework for AUVs that enhances robustness against sensor noise and DVL outages.
• To improve the accuracy and continuity of AUV navigation during periods of DVL unavailability.

Main Methods:

• Proposed an integrated SINS/DVL/PS navigation framework combining an Adaptive Huber and Sage-Husa Extended Kalman Filter (AHR-EKF) with a Nonlinear AutoRegressive with eXogenous inputs (NARX)-based velocity prediction model.
• The AHR-EKF addresses time-varying noise and outliers, while the NARX model predicts velocity during DVL outages using propeller speed and navigation system data.
• Validated the framework through simulations and sea trials, comparing its performance against traditional methods.

Main Results:

• The V-NARX method significantly reduced positioning root mean square errors during simulated DVL outages compared to the V-RPM method (e.g., 8.397 m vs. 24.699 m East).
• Sea trials demonstrated substantial improvements with V-NARX, achieving 22.1% and 58.2% reductions in East and North RMS errors, respectively, during DVL outages.
• The proposed method effectively suppressed inertial navigation system (INS) error accumulation and maintained trajectory continuity during DVL outages.

Conclusions:

• The integrated AHR-EKF and NARX-based navigation framework offers a practical and reliable solution for continuous AUV navigation during DVL outages.
• This approach significantly enhances the emergency navigation capabilities of AUVs in challenging underwater conditions.
• While not matching normal DVL accuracy, the method provides a crucial fallback for maintaining operational continuity.