I am the student project lead for the Bobcat-1 Cubesat (launching next month!). @DL4PD shared the LSF Polaris project (https://gitlab.com/librespacefoundation/polaris/polaris) with me, and I wanted to share how we plan to use it to aid our mission goals.
For those who are unfamiliar with Polaris, it is a tool that takes decoded frames from SatNOGS observations and uses machine learning to find links and correlation between telemetry fields and subsystems. For example, see the demo site with LightSail-2 data here: https://deepchaos.space/
Bobcat-1’s mission revolves around GNSS experiments using a high accuracy GNSS receiver and a custom GPS+Galileo SDR receiver. Obviously, anything to do with navigation satellites has a lot to do with space weather, so one of the things we’re really excited to do is postprocess our downlinked GNSS data and feed it to Polaris along with our SatNOGS frames and some space weather data, and see what links we find. This will hopefully lead to some interesting results, stay tuned for that!
In general though, we plan to run the Polaris fetch and learn scripts regularly on our groundstation, perhaps every 2-3 days, and use it as a tool along with our SatNOGS dashboard to better understand the actual operating characteristics of our Cubesat in LEO and hopefully spot potential issues before they impact our experiments. Although we have a good understanding of how our hardware and software works, it will be helpful to have Polaris as another tool in our toolbox to understand how the cubesat is behaving in the space environment.