Improved Target Trajectory Reconstruction in HFSWRs Using a DBSCAN Clustering Algorithm

In this paper we propose the enhanced method for target trajectory reconstruction in High Frequency Surface Wave Radars using DBSCAN clustering algorithm. The goal is to improve the accuracy and reliability of reconstructed trajectories by clustering detections obtained by radar’s high-resolution primary signal processing. Although the paper does not focus directly on the trajectory creation from obtained clusters, it demonstrates how clustering can be successfully applied along known vessel trajectories which are verified by Automatic Identification System data. The experimentally obtained results confirm the validity of the proposed method which has significant potentials for further performance improvement of maritime vessel tracking.