Satellites are rapidly gaining advanced processing capabilities that can be used to execute machine learning algorithms and, thus, result in having a certain degree of intelligence at the individual satellites. This intelligence enables the use of advanced network resource management mechanisms.

One of the essential networking tasks is routing, where the goal is to find the optimal route for the data from source to destination, when these are located at different areas around the Earth.

This project focuses on using the intelligence at the satellites to train models for routing in large non-geostationary orbit (NGSO) satellite constellations. Due to the movement of the NGSO satellites, the routing algorithm must be capable of learning the optimal route for a nearly-infinite number of different positions of the satellites in the constellation with respect to the Earth. Hence, the goal is to design deep reinforcement learning mechanisms to allow the satellites to perform autonomous and optimal routing decisions.

ABOUT SATNEX V

SatNEx V is a network of satellite experts dealing with the scenario of data processing and communication in a hybrid infrastructure of GEO and LEO satellites.

SatNEx V is a European Space Agency project within the framework Technology Development Programme. The project involves a range of university partners from all across Europe, including Aalborg University.