Adaptive DBSCAN-Based Probabilistic Data Association Filter for Maritime Object Tracking Using RADAR

This study introduces a robust methodology for maritime object tracking using RADAR data without relying on automatic identification system and vision sensors. By employing adaptive density-based spatial clustering of applications with noise (ADBSCAN) clustering for target initialization and probabilistic data association filter (PDAF) for tracking, the framework accurately distinguishes dynamic ships from static clutter. The integration of ADBSCAN enables adaptive clustering to handle varying RADAR data densities, while PDAF ensures robust tracking in noisy conditions. Experiments demonstrate reliable target detection and tracking performance, as well as resilience to data loss, highlighting its potential for autonomous navigation in complex maritime environments.