Decentralised Semi-supervised Onboard Learning for Scene Classification in Low-Earth Orbit
This work addresses the problem of reducing communication and operational costs for satellite operators by enabling efficient onboard learning, though it appears incremental as it applies existing semi-supervised and federated learning methods to a new domain with specific constraints.
The paper tackled onboard machine learning for scene classification in satellite constellations using semi-supervised learning under operational constraints like temperature and power limits, achieving convergence to around 91% accuracy on the EuroSAT RGB dataset within a one-day mission timeframe.
Onboard machine learning on the latest satellite hardware offers the potential for significant savings in communication and operational costs. We showcase the training of a machine learning model on a satellite constellation for scene classification using semi-supervised learning while accounting for operational constraints such as temperature and limited power budgets based on satellite processor benchmarks of the neural network. We evaluate mission scenarios employing both decentralised and federated learning approaches. All scenarios achieve convergence to high accuracy (around 91% on EuroSAT RGB dataset) within a one-day mission timeframe.