Online Learning for Dynamic Constellation Topologies

arXiv cs.LG / 3/30/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper studies how to configure dynamically changing satellite network topologies using an online learning framework that accounts for orbital motion and satellite maneuvering.
It avoids relying on fixed structural assumptions such as known orbital planes, which may be invalidated by maneuvers, making the approach more robust to real-world dynamics.
The authors show empirically that their online formulation achieves performance comparable to state-of-the-art offline topology configuration methods.
They demonstrate that the method can be adapted to constrained online learning settings, highlighting a trade-off between per-iteration computational complexity and convergence to a final strategy.

Abstract

The use of satellite networks has increased significantly in recent years due to their advantages over purely terrestrial systems, such as higher availability and coverage. However, to effectively provide these services, satellite networks must cope with the continuous orbital movement and maneuvering of their nodes and the impact on the network's topology. In this work, we address the problem of (dynamic) network topology configuration under the online learning framework. As a byproduct, our approach does not assume structure about the network, such as known orbital planes (that could be violated by maneuvering satellites). We empirically demonstrate that our problem formulation matches the performance of state-of-the-art offline methods. Importantly, we demonstrate that our approach is amenable to constrained online learning, exhibiting a trade-off between computational complexity per iteration and convergence to a final strategy.