Welcome to Polaris’s documentation!¶
Polaris is an open source (LGPLv3) project that analyzes satellite telemetry using machine learning. It is a project of the Libre Space Foundation.
Want to get started quickly? Have a look at Getting started with Polaris.
Polaris has a number of different commands:
polaris fetchwill download and normalize satellite telemetry from the SatNOGS network (or you can import your own).
polaris learnwill analyze the telemetry, produce a model of the connections between telemetry components, and save a dependency graph for visualization.
polaris vizis an interactive, browser-based 3D visualization of that dependency graph.
polaris convertwill convert graph output from
polaris learnto another file format (like
polaris behavewill detect anomalies in telemetry data and produce a json report of all the data and anomaly produced.
polaris reportwill show the results obtained by
polaris behavein web-based interactive formats like anomaly visualisation, individual telemetry comparison etc.
For more details about how Polaris works, see Overview, or watch our presentation at the 2020 Cubesat Developer’s Workshop.
Would you like to join the community? We’d love to get to know you! Join our chat room here and introduce yourself.
If you’d like to work with us through the Google Summer of Code, have a look at our notes for GSoC applicants.
For other presentations on Polaris, see Presentations on Polaris: Past and Future.