Unmanned aerial vehicle classification and detection system based on deep learning, internet of military things, and PID control system

Indonesia is an archipelagic country situated between two continents and two oceans. With numerous islands, it is rich in natural resources but faces various military and non-military threats. One significant threat to maritime nations like Indonesia is from the air, which includes direct attacks from manned and unmanned aircraft and using aerial vehicles for intelligence and surveillance. The primary weapon system is crucial for national defense against such threats. Therefore, developing defense equipment in Indonesia must align with technological advancements to ensure quick and efficient operation. This research focuses on creating a classification and reconnaissance system for flying vehicles to enhance air defense capabilities. In the surveillance system, two servos are used for yaw and pitch axes, controlled by a Proportional, Integrative, and Derivative (PID) system. This PID control significantly improves servo movement both dynamically and statically. The system sends notifications via Telegram for monitoring, with an average FPS of 9.6. Flask is used for the website interface, averaging 6.8 FPS, and MIT App Inventor is used for the smartphone interface, averaging 7.6 FPS. This flying vehicle classification and reconnaissance system enhances Indonesia's air defense, utilizing YOLOv8 for classification, PID control for servo movements, and integrated notifications and interfaces for both web and smartphones.