A Robust INS/USBL/DVL Integrated Navigation Algorithm Using Graph Optimization.

Area of Science:

• Robotics and Autonomous Systems
• Navigation and Positioning
• Signal Processing

Background:

• Autonomous Underwater Vehicles (AUVs) rely on multiple sensors like INS, USBL, and DVL for navigation.
• Accurate underwater navigation requires effective multi-source information fusion.
• Environmental factors can introduce outliers and non-Gaussian noise in sensor data.

Purpose of the Study:

• To propose a robust factor graph optimization (FGO) based fusion scheme for AUV navigation.
• To address challenges posed by sensor outliers and non-Gaussian noise in underwater environments.
• To enhance the precision and reliability of AUV multi-sensor integrated navigation.

Main Methods:

• Implemented a factor graph optimization (FGO) framework for multi-sensor data fusion.
• Developed a robust integrated navigation algorithm incorporating the maximum correntropy criterion.
• Combined FGO with robust estimation techniques to handle outliers and non-Gaussian noise.

Main Results:

• The proposed FGO scheme demonstrated superior performance and robustness compared to traditional Kalman filters.
• Numerical simulations and field tests validated the effectiveness of the FGO approach.
• Achieved significant improvements in positioning accuracy: 5.3% (East), 9.1% (North), and 5.1% (Height).

Conclusions:

• The robust FGO scheme effectively fuses multi-sensor data for AUVs, even with environmental noise.
• This method provides more accurate underwater navigation and positioning than conventional techniques.
• The approach offers a reliable solution for high-precision underwater multi-sensor navigation.