Autonomous Air Defense System Supported by Embedded Image Processing
2025 Innovations in Intelligent Systems and Applications Conference, ASYU 2025, Bursa, Türkiye, 10 - 12 Eylül 2025, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/asyu67174.2025.11208382
- Basıldığı Şehir: Bursa
- Basıldığı Ülke: Türkiye
- Anahtar Kelimeler: autonomous air defense systems, autonomous decision-making, convolutional neural networks, deep learning, embedded image processing, threat detection, unmanned aerial vehicles
- Ankara Yıldırım Beyazıt Üniversitesi Adresli: Evet
Özet
The rapid evolution of autonomous technologies has transformed air defense systems, where embedded image processing enhances threat detection and response speed and accuracy. This paper explores integrating deep learning, sensor fusion, and autonomous decision-making into Autonomous Air Defense Systems (AADS) to detect, track, and neutralize airborne threats like missiles and drones with minimal human input. CNNs process real-time sensor data for target classification and tracking, while sensor fusion improves accuracy by combining multiple sensor inputs. AI-driven decision-making further streamlines defense strategies, reducing response times and complexity. This paper reviews the latest research and technological advancements in the field, discussing key innovations, challenges, and future directions in embedded image processing for autonomous air defense. Through a detailed analysis of recent literature, we aim to provide a comprehensive understanding of the current state of AADS, highlighting both the opportunities and limitations of integrating advanced image processing techniques in autonomous military systems. The system is positioned as a robust prototype for a full-fledged air defense system. In the scope of the study, the mechanical infrastructure was meticulously designed for high-precision targeting and operational robustness. The electronic hardware, integrating a Raspberry Pi 5 and an NVIDIA RTX 4060-equipped computer, enables YOLOv11n target detection and Channel and Spatial Reliability Tracking (CSRT). The software architecture, developed modularly with Python and relevant libraries, incorporates AI-powered decision-making mechanisms such as intelligent prioritization, adaptive tracking, and autonomous firing authorization. Comprehensive safety protocols, including human detection and emergency stop integration, have also been implemented. Completed simulations and physical tests have validated the system's performance and reliability.