Decentralised Semi-supervised Onboard Learning for Scene Classification in Low-Earth Orbit

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.

[1]  Gabriele Meoni,et al.  PAseos Simulates the Environment for Operating multiple Spacecraft , 2023, ArXiv.

[2]  Aizaz U. Chaudhry,et al.  Laser Inter-Satellite Link Setup Delay: Quantification, Impact, and Tolerable Value , 2023, IEEE Wireless Communications and Networking Conference.

[3]  Israel Leyva Mayorga,et al.  Federated Learning in Satellite Constellations , 2022, IEEE Network.

[4]  Armin Dekorsy,et al.  On-Board Federated Learning for Dense LEO Constellations , 2021, ICC 2022 - IEEE International Conference on Communications.

[5]  Gabriele Meoni,et al.  MSMatch: Semisupervised Multispectral Scene Classification With Few Labels , 2021, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[6]  Quoc V. Le,et al.  EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks , 2019, ICML.

[7]  Indranil Gupta,et al.  Asynchronous Federated Optimization , 2019, ArXiv.

[8]  Andreas Dengel,et al.  Introducing Eurosat: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification , 2018, IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium.

[9]  Aubrey K. Dunne,et al.  The Φ-Sat-1 Mission: The First On-Board Deep Neural Network Demonstrator for Satellite Earth Observation , 2022, IEEE Transactions on Geoscience and Remote Sensing.