Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

High-level data fusion component for drone classification and decision support in counter UAV

: Sander, Jennifer; Kuwertz, Achim; Mühlenberg, Dirk; Müller, Wilmuth


Suresh, R. ; Society of Photo-Optical Instrumentation Engineers -SPIE-, Bellingham/Wash.:
Open Architecture/Open Business Model Net-Centric Systems and Defense Transformation 2018 : 7-19 April 2018, Orlando, Florida, United States
Bellingham, WA: SPIE, 2018 (Proceedings of SPIE 10651)
ISBN: 978-1-5106-1814-5
ISBN: 978-1-5106-1813-8
Paper 106510F, 10 pp.
Conference "Open Architecture/Open Business Model Net-Centric Systems and Defense Transformation" <23, 2018, Orlando/Fla.>
Conference "Defense and Security" <2018, Orlando/Fla.>
Conference "Defense and Commercial Sensing" (DCS) <2018, Orlando/Fla.>
Conference Paper
Fraunhofer IOSB ()
counter UAV; MODEAS; drone classification; knowledge-based approach; high-level data fusion; decision support; threat assessment; behavior analysis; object database; STANAG 4559

Today, drone technology has been made available around the world. Anyone can purchase a drone from an online retailer. Government agencies and military are seeing a rise in drones used for terrorism, destruction and espionage. The emergence of threats caused by unfriendly or hostile drones requires proactive drone detection in order to decide on appropriate defensive actions. In this contribution, a high-level data fusion component for drone classification is presented. The high-level data fusion component is part of our counter UAV system MODEAS including decision support. The component provides well-defined interfaces which allow it to be integrated also into other counter UAV systems. The aim of the high-level data fusion component is to support an operator in his decision making by providing detailed information about detected drones together with assigned threat levels. To identify a detected and tracked drone with sufficient detail, a knowledge-based classification is performed, based on background knowledge like drone model specifications. By fusing the knowledge-based classification results with prior results of a sensor-based classification, the overall classification is improved. The fusion results, in addition to kinematic data, also contain specific capabilities of the respective drone like its maximum payload, endurance, and speed as well as recorded incidents with similar drones or their typical (commercial) usage, if known. Based on these fusion results, a threat analysis is performed. The component’s output then is a ranked list of dossiers for the most probable types of drones with regard to the observation data and their assigned threat levels.