Abstract
The rapid expansion of satellite constellations in near-Earth orbits presentssignificant challenges in satellite network management, requiring innovativeapproaches for efficient, scalable, and resilient operations. This paperexplores the role of Artificial Intelligence (AI) in optimizing the operationof satellite mega-constellations, drawing from the ConstellAI project funded bythe European Space Agency (ESA). A consortium comprising GMV GmbH, SaarlandUniversity, and Thales Alenia Space collaborates to develop AI-drivenalgorithms and demonstrates their effectiveness over traditional methods fortwo crucial operational challenges: data routing and resource allocation. Inthe routing use case, Reinforcement Learning (RL) is used to improve theend-to-end latency by learning from historical queuing latency, outperformingclassical shortest path algorithms. For resource allocation, RL optimizes thescheduling of tasks across constellations, focussing on efficiently usinglimited resources such as battery and memory. Both use cases were tested formultiple satellite constellation configurations and operational scenarios,resembling the real-life spacecraft operations of communications and Earthobservation satellites. This research demonstrates that RL not only competeswith classical approaches but also offers enhanced flexibility, scalability,and generalizability in decision-making processes, which is crucial for theautonomous and intelligent management of satellite fleets. The findings of thisactivity suggest that AI can fundamentally alter the landscape of satelliteconstellation management by providing more adaptive, robust, and cost-effectivesolutions.