Development and Performance Evaluation of a Neuro-Based Software for Modelling of Orbital Properties of Satellites

Abstract

Satellite communication systems have become a revolutionary communication technology as a result of their ability to cover large areas and transmit information over otherwise difficult terrain. However, the number of such satellites being launched into space is increasing drastically, and techniques to accurately model the motion and position of these satellites have become a necessity. In this work, we developed neuro-based software for modelling the orbital properties of satellites and analyzed the performance thereof. This was achieved by optimizing an artificial neural network trained by the Gradient Descent with Momentum (GDM) algorithm, and linking it with the graphical user interface of our proposed software. The overall performance of the system when evaluated with different datasets shows that the accuracy and speed were satisfactory. The results, therefore, stand as a good tool for educational and training purposes, highlighting the growing impact of the use of expert systems in aviation, remote sensing, and satellite communication.