Taxonomy for Resident Space Objects in LEO: A Deep Learning Approach

[In the Journal]

Authors:

• Neuraspace, Portugal, {marta.guimaraes, chiara.manfletti}@neuraspace.com
• FCT-UNL, Portugal, claudia.soares@fct.unl.pt

Having a well-established taxonomy of Resident Space Objects (RSOs) enables assigning objects to specific categories, leading to better tracking services, thereby achieving a better understanding of the RSOs. This will in turn help design efficient and effective strategies for Space Traffic Management.

Download the paper to learn more about:

• The increasing number of RSOs and its implications for Space Sustainability
• Data sources used
• Deep learning model using autoencoder architecture
• Latent space visualization and clustering
• Extracting insights from clusters using the Decision Tree-based model, and
• Taxonomy