Mmdalga Radar Kullanarak Drone Sürülerinin ISAR Görüntülenmesi
No Thumbnail Available
Date
2025
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Open Access Color
OpenAIRE Downloads
OpenAIRE Views
Abstract
Teknolojide drone ve drone sürülerinin kullanımının artmasıyla birlikte, anti-drone teknolojilerinin kullanımı önemli ölçüde artmıştır. Ancak, sınırlı görüş alanına sahip senaryolarda drone ve drone sürülerinin tespiti literatürde kalıcı bir zorluk olmaya devam etmektedir. Bu tez, milimetre dalga (mmWave) frekans bantlarında yeniden oluşturulan drone sürülerinin Ters Sentetik Açıklıklı Radar (ISAR) görüntülerinin, oluşumlarına, boyutlarına ve yük yapılandırmalarına göre sınıflandırılmasına odaklanmaktadır. Drone sürülerinin ISAR görüntüleri, ANSYS Yüksek Frekanslı Yapısal Simülatör (HFSS) elektromanyetik simülasyon yazılımı kullanılarak üretilmiştir. Sürü yapıları, quadcopter dronlar kullanılarak modellenmiş ve oluşum tipleri, çizgi, çarpı, kare ve üçgen gibi temel geometrik şekillerle tasarlanmıştır. Sürülerdeki dronlar, orta, küçük ve mini olmak üzere üç boyutta kategorize edilmiştir. Ek olarak, yük dronları sürü yapılandırmalarına dahil edilmiştir. Yüksek çözünürlüklü ISAR görüntüleri elde etmek için radar ve simülasyon parametreleri optimize edilmiştir. Veri setini genişletmek için, ISAR görüntüleri çeşitli bakış açılarında (0° ila 350° arasında 10° artışlarla) oluşturulmuştur. ISAR görüntüleri kullanılarak sürü oluşumu tiplerinin belirlenmesi, görüntü tanıma aşamasında bir Evrişimsel Sinir Ağı (CNN) aracılığıyla gerçekleştirildi. Bunu takiben, nesne algılama aşamasında Sadece Bir Kez Bak (YOLO) algoritması kullanılarak drone boyutu ve yük tespiti gerçekleştirildi. Bu tezde elde edilen sonuçlar oldukça ümit vericidir. Genişletilmiş bir veri seti ve tespit algoritması sunarak, bu çalışma literatüre önemli katkıda bulunmaktadır.
With the increasing use of drones and drone swarms in technology, the importance of anti-drone technologies has grown significantly. However, detecting drones and drone swarms in scenarios with limited fields of view remains a persistent challenge in the literature. This thesis focuses on the classification of Inverse Synthetic Aperture Radar (ISAR) images of drone swarms reconstructed at millimeter-wave (mmWave) frequency bands based on their formation, size, and payload configurations. ISAR images of drone swarms were generated using the ANSYS High-Frequency Simulation Software (HFSS) electromagnetic simulation software. Swarm structures were modeled using quadcopter drones, and the formation types were designed with basic geometric shapes, such as line, cross, square, and triangle. The drones in the swarms were categorized into three sizes: medium, small, and mini. Additionally, payload drones were included in the swarm configurations. Radar and simulation parameters were optimized to obtain high-resolution ISAR images. To expand the dataset, ISAR images were reconstructed at various look angles (from 0° to 350° in 10° increments). The determination of swarm formation types using ISAR images was carried out through a Convolutional Neural Network (CNN) in the image recognition phase. The detection of drone size and payload configurations was conducted using the You Only Look Once (YOLO) algorithm in the object detection phase. The results achieved in this thesis are highly promising. By presenting an extended dataset and detection algorithm, this work contributes significantly to the literature and advances the field of drone swarm detection.
With the increasing use of drones and drone swarms in technology, the importance of anti-drone technologies has grown significantly. However, detecting drones and drone swarms in scenarios with limited fields of view remains a persistent challenge in the literature. This thesis focuses on the classification of Inverse Synthetic Aperture Radar (ISAR) images of drone swarms reconstructed at millimeter-wave (mmWave) frequency bands based on their formation, size, and payload configurations. ISAR images of drone swarms were generated using the ANSYS High-Frequency Simulation Software (HFSS) electromagnetic simulation software. Swarm structures were modeled using quadcopter drones, and the formation types were designed with basic geometric shapes, such as line, cross, square, and triangle. The drones in the swarms were categorized into three sizes: medium, small, and mini. Additionally, payload drones were included in the swarm configurations. Radar and simulation parameters were optimized to obtain high-resolution ISAR images. To expand the dataset, ISAR images were reconstructed at various look angles (from 0° to 350° in 10° increments). The determination of swarm formation types using ISAR images was carried out through a Convolutional Neural Network (CNN) in the image recognition phase. The detection of drone size and payload configurations was conducted using the You Only Look Once (YOLO) algorithm in the object detection phase. The results achieved in this thesis are highly promising. By presenting an extended dataset and detection algorithm, this work contributes significantly to the literature and advances the field of drone swarm detection.
Description
Keywords
Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
Turkish CoHE Thesis Center URL
Fields of Science
Citation
WoS Q
Scopus Q
Source
Volume
Issue
Start Page
End Page
72