Satellites state vectors refinement based on international laser ranging system using machine and deep learning

Abstract

With the increasing number of space objects in near-Earth space, the necessity of high-precision determination of space objects state vectors, as well as its classification by size, velocity, and potential danger to active satellites and space stations is becoming increasingly important for space flight safety services. In case of necessity of taking decisions of satellites orbit corrections and avoiding space emergency situations in real time mode artificial intelligence services could be used. The results proposed in this study show that machine and deep learning models can significantly improve the accuracy of determining the space objects state vector for classical numerical models and space catalogs, that is very essential task for space flights safety. The parameters of the Two-Line-Elements catalog and the model of it convertation to state vector are considered as input data to process, International Laser Ranging Service data from ground stations is considered as the ground truth measurements. The methodology considered here can be applied to any artificial space objects with various orbit parameters, thus, helps to provide space flights safety assurance.