Analytical and deep learning approaches for interference detection in LEO satellite constellations

The current deployment of mega-constellations of telecommunication satellites in low Earth orbit is paving the way for global broadband connectivity and offers prospects for the Internet of Things, Earth observation and defence applications. However, the significant increase in the number of satellites in orbit, operating in the same frequency bands at these altitudes increases interference. In this paper, one analytical algorithm and one deep learning algorithm for low Earth orbit satellite interference detection tasks are introduced. These solutions are developed and tested on realistic synthetic data, using a customized orbital simulator. The proposed methods demonstrate promising performance, combining low computational complexity with high detection accuracy compared to state-of-the-art methods.