Aerospace Instrument-Making Annotation << Back Clustering of Dynamic Object Observations Gathered by Several Sources VIA the DBSCAN Method With the Bhattacharya Metric Zhegalov S.N., Maslennikov A.L. The problem of identifying dynamic object observations gathered from multiple sources is considered. An observation is considered as a vector of the dynamic object's coordinates and velocities, determined by the source relative to its coordinate system. Identification is performed using clustering via the DBSCAN algorithm. It is proposed to use the Bhattacharyya metric in the DBSCAN algorithm, which considers the probabilistic characteristics of coordinates and velocities measurement errors, and a modification of this metric, which considers the importance of each element of the observation vector in the clustering process. Performed computational experiments demonstrated a successful solution to the identifi cation problem, including the case of intersecting trajectories of dynamic objects. The use of the Bhattacharyya metric and its modifications improve the quality of clustering compared to the Euclidean metric, which is the basis for DBSCAN. Keywords: coordinate observations, dynamic objects, clustering, DBSCAN, Bhattacharyya distance. DOI: 10.25791/aviakosmos.3.2026.1543 Pp. 25-36.