A Sequential Kalman-Newton-KM Framework for AIS and Radar Data Fusion in Restricted Inland Waterways

Area of Science:

• Maritime technology
• Sensor fusion
• Data science

Background:

• Real-time vessel monitoring in restricted waterways is crucial for traffic management.
• Existing systems face challenges due to data sparsity from Automatic Identification System (AIS) and lack of identity in radar data.

Purpose of the Study:

• To develop a novel data fusion framework integrating AIS and radar data for enhanced real-time vessel monitoring.
• To improve vessel tracking accuracy and situational awareness in inland restricted waterways.

Main Methods:

• Developed a framework combining Kalman filtering and Newton interpolation (K-N) for high-resolution AIS resampling.
• Employed the Kuhn-Munkres (KM) algorithm for optimal data association, treating it as a global optimization problem.
• Utilized virtual point augmentation to handle data imbalance between heterogeneous sensors.

Main Results:

• Achieved high matching accuracy: 94.2% in low-density and 80.1% in high-traffic scenarios.
• Demonstrated computational efficiency suitable for real-time deployment.
• Showcased consistent performance across various waterway geometries, with slight variations in curved versus straight channels.

Conclusions:

• The proposed framework effectively fuses AIS and radar data for accurate and reliable vessel tracking.
• Enhanced situational awareness is provided to waterway authorities for improved traffic management.
• The system offers a robust solution for real-time vessel monitoring in challenging inland waterway environments.