Satellite-as-a-Service for Onboard Machine Learning
I got to present one of our neat projects at this year’s IEEE AeroConf, developed by my consultancy Tanagra Space in collaboration with VisionSpace and ESA-ESOC:
Labrèche, G., & Alvarez, C. G. (2023). SaaSyML: Software as a Service for Machine Learning On-board the OPS-SAT Spacecraft. In 2023 IEEE Aerospace Conference, Big Sky, MT, USA, 2023, pp. 1-9, doi: 10.1109/AERO55745.2023.10115531.
The European Space Agency’s OPS-SAT space lab is a versatile platform that enables software engineering using modern web 2.0 best practices. The SaaSyML app provides experimenters with a RESTful API for on-board Machine Learning applications. It’s Space Platform as-a-Service (SPaaS)! It implements a publish-subscribe pattern that uses on-board data as training data and can run in-flight inferences with the trained models.
A massive thank you to the OPS-SAT Mission Control Team for all their support in making this experiment a success and a big shout-out to the AeroConf event organizers! Here’s the source code.
A version of this post was first published on LinkedIn.