Correntropy Based Divided Difference Filtering for the Positioning of Ships

Area of Science:

• Signal Processing
• Navigation Systems
• Robust Estimation

Background:

• Divided difference filters are effective but sensitive to non-Gaussian noise.
• Ship positioning often relies on integrating Dead Reckoning (DR) with Global Positioning System (GPS).
• Heavy-tailed noise significantly degrades the performance of conventional filters in navigation.

Purpose of the Study:

• To develop robust filtering algorithms for improved navigation accuracy.
• To enhance the resilience of ship positioning systems against non-Gaussian noise.
• To leverage correntropy for robust estimation in dynamic systems.

Main Methods:

• Proposed first and second-order divided difference filtering algorithms utilizing correntropy.
• Applied the developed filters to a combined Dead Reckoning (DR) and Global Positioning System (GPS) ship positioning framework.
• Evaluated filter performance using experimental data and comparative analysis.

Main Results:

• The correntropy-based filters demonstrated superior robustness against heavy-tailed and non-Gaussian noise.
• The proposed algorithms significantly improved the accuracy and reliability of ship positioning.
• Experimental results confirmed the enhanced performance compared to existing methods.

Conclusions:

• Correntropy-based divided difference filters provide a robust solution for navigation under challenging noise conditions.
• The integration of these filters with DR-GPS enhances ship positioning performance.
• This approach offers a promising direction for reliable autonomous and assisted navigation systems.