Megaconstellations: is it possible to achieve the same performance using fewer satellites?

The new frontier of the Internet will be satellite connections through which it will be possible to provide broadband Internet with global coverage in the coming years. This will be made possible by constructing mega-constellations of satellites operating in low Earth orbit (LEO) or very low Earth orbit (VLEO), ensuring very competitive latency times far shorter than the latency times experienced with current geostationary constellations. Among the constellations that will be active in the coming decades, Starlink is the one that is currently attracting particular attention as its final configuration has a huge number of satellites and its deployment is already underway. This high number of satellites greatly concerns the scientific community mainly because it dramatically increases the possibility that the Kessler syndrome may be realized, and thus space will become inaccessible for the next decades or centuries. For this very reason, a study aimed at understanding the reasons behind SpaceX’s design of this constellation becomes necessary, and it would be interesting also to see if there is a way to lower the number of satellites required. The following study, therefore, looks at the final configuration of the constellation in which 42,000 satellites distributed over 16 shells capable of communicating via laser technology are involved. Given the orbital parameters of each shell, it was necessary to propagate their orbits using the Walker pattern. However, the first considerable difficulty was the identification of the F parameter, which is used to determine the phasing between the satellites in the different orbital planes and consequently serves to avoid a collision between satellites. After that, knowing the position of the satellites in time, to estimate the optimal latency times, Dykstra’s algorithm was applied, which allowed the identification of the best path in the network of satellites. This provided insight into the potential of the constellation in terms of coverage, latency and connection stability, and allowed for further investigation by studying the effects of reducing the number of satellites on the performance of the system. There are few accessible data, and those used in this thesis have been released by the Federal Communications Commission and Starlink itself on its website. Consequently, it was necessary to do a simplified study in which possible traffic congestion was not taken into account. However, this gave us interesting results; for example, it was shown how, in the case of the most numerous shells, even a 70% reduction of satellites does not affect coverage and minimum latency times. In general, all the results obtained are interesting and serve as a stimulus to encourage subsequent, more in-depth studies in which other aspects unknown today, such as data traffic, are evaluated.