Polaris is a project brought together by a team of contributors from around the world and Libre Space Foundation.

Polaris is an open-source tool used for the exploration and analysis of satellite telemetry data by building machine learning models. It is built on Python and uses the telemetry data provided by the SatNOGS network. The models are used for better understanding and predicting more accurately the behaviour of a satellite. By analysing telemetry and other data sources and converting the data into useful information for spacecraft operators.

Polaris brings impressive graph visualization of the telemetry parameters dependencies, helping spacecraft operators to navigate and to drive their investigations when anomalies occur. In the coming months, the team will use deep learning to extract the link between the behaviours of these different parameters. This is a step toward autonomous explanations of the changes in behaviour while some of these changes are actual anomalies. This will help operators with autonomous diagnostics.

Polaris aims at bringing AI and Machine Learning to Satellites and it aspires to “pave the way toward open-source, autonomous Satellite operations for missions of all scales anywhere in the solar system”.

If you want to find out more about this project check out the open-source repositories.
Find documentation, content and demos on the Polaris website.
Join the open discussions about the project and its evolution on the dedicated channel. You can also check a presentation of the project and all the latest updates: