Navigation science students in Beni Suef develop a satellite collision avoidance subsystem by detecting space debris

Media Center
A research team of students of the Faculty of Navigation Sciences and Space Technology under the auspices of Dr. Mansour Hassan, President of Beni Suef University, Under the supervision of Dr. Osama Shalabiya, Dean of the College, the development of an AI Vision computer sub-system to avoid satellite collisions through the discovery of space debris, by providing satellites with capabilities to detect and analyze space debris in real time, which contributes effectively to public safety and the sustainability of space operations
The President of the University expressed his happiness with the level of graduation projects submitted by students of the first batch of the College of Navigation Sciences and Space Technology, and thanked all those in charge of the college and the students, praising those ideas and research projects that can be used in the field of navigation and space.
Dr. Osama Shalabiya added that the graduation project was carried out under the supervision of Dr. Muhammad Al-Fran, Lecturer of the Department of Astronautics,
The project is based on the development and implementation of a powerful computer vision artificial intelligence system that enables satellites to detect space debris and independently maneuver to avoid potential collisions. The project aims to design and integrate a computer vision-based system capable of real-time detection, classification, and trajectory analysis of space debris by using of deep learning algorithms. The collision of satellites with space debris consider a major threat to the sustainability of space operations to reduce these risks, the students implemented a new subsystem to enhance satellite collision avoidance capabilities by incorporating computer vision artificial intelligence (AI).
The students trained their AI model, using a diverse dataset of space debris imagery, to identify potential collision risks and accurately track them to ensure feasibility and efficiency, optimizing the AI system to operate on-board the satellite with minimal computational resources. The project methodology involved collecting a comprehensive dataset of space debris imagery, covering a wide range of debris types and sizes, through the use of convolutional neural networks (CNNs) and transfer learning techniques
The subsystem was integrated into the existing satellite system, with a focus on seamless compatibility and minimal impact on its overall operations. Through rigorous testing, the student-designed system showed outstanding performance in detecting space debris in real time. The results indicated: Average detection accuracy of over 95% for various sizes and types of debris, This enables the timely assessment of collision risks. The AI Vision computer subsystem provides reliable data for collision avoidance decisions, enabling satellites to independently adjust their orbits or perform evasive maneuvers