An Application of Online Learning to Spacecraft Memory Dump Optimization

TL;DR

This paper demonstrates how online learning algorithms can significantly improve spacecraft memory dump efficiency, achieving over 60% performance gains on real satellite data.

Contribution

It introduces a novel application of online learning with expert advice to optimize spacecraft memory management in real-world scenarios.

Findings

Over 60% performance improvement with online learning

Successful application on Sentinel-6 satellite data

Lightweight Follow-The-Leader algorithm outperforms traditional methods

Abstract

In this paper, we present a real-world application of online learning with expert advice to the field of Space Operations, testing our theory on real-life data coming from the Copernicus Sentinel-6 satellite. We show that in Spacecraft Memory Dump Optimization, a lightweight Follow-The-Leader algorithm leads to an increase in performance of over $60\%$ when compared to traditional techniques.