| Avoiding Satellite Collision Approach Based on Recurrent Neural Network |
| Paper ID : 1111-ISCHU |
| Authors |
|
Alaa osama awad *1, Sara Abdelghafar2, Mourad Raafat3, Ashraf Darwish4, Aboul Ella Hassanien5 1Department of Mathematics, Faculty of Science, Helwan University, Cairo, Egypt 2Faculty of science,Al Azhar university,Cairo,Egypt 3Faculty of scince,Helwan university, Cairo,Egypt 4Faculty of scince, Helwan university, Cairo,Egypt 5Faculty of Computers and Artificial Intelligence, Cairo university, Giza,Egypt |
| Abstract |
| Active collision avoidance is becoming a significant responsibility in space operations, with hundreds of alerts issued each week for a satellite in Low Earth Orbit correlating to close interactions with other space objects. As a result, orbit prediction plays a key and essential part in satellite monitoring and tracking control, The first two satellites collision occurred in February 2009 collision of a U.S. Iridium communications satellite and a Russian Cosmos 2251 communication satellite and this collision increased the number of space debris by approximately 13%. In this research working on avoiding satellite collisions by forecasting and predicting satellite position and velocity using data from Kaggle, which provides real and simulated data for six hundred satellites with satellite positions (x, y, z) and velocity (Vx, Vy, Vz). using Long Short-Term Memory (LSTM), a Recurrent Neural Network (RNN) architecture, then compare deep learning predicting and forecasting data with Kaggle simulation data. The experimental findings suggest that the proposed technique can accurately estimate satellite orbits. |
| Keywords |
| Satellite Orbit, Orbit Prediction, Artificial Intelligence, Deep Learning, Recurrent Neural Network, Long Short-Term Memory, time series forecasting. |
| Status: Abstract Accepted (Oral Presentation) |