For the second year in a row, Libre Space Foundation was selected as a mentor organisation for the Google Summer of Code initiative. The application period has closed and the results are in, and so it is with great excitement that we announce the two projects we will be mentoring over the next few months.
The first project titled “Deep learning for Cubesat Behavior Segmentation with Collection of Contextual Information” will be working on the Polaris codebase. The project aims at supporting spacecraft operators by predicting the behaviour of their satellites and linking it to various data sources. There is a data challenge in collecting and sometimes in converting into time series. This data collection phase will allow for better information when understanding and estimating the behaviour of a spacecraft. External sources of data, namely, orbit propagation, solar and magnetic events, and various elements of space weather, will be some of the external sources providing the data needed. The machine learning approach employed for Polaris will transform these data sources into learning features so that a spacecraft’s behaviour is not only predicted but also explained by the “machine”. Deep learning means that the project is exploring the usage of different neural network architectures of several layers. The project is undertaken in close collaboration with the amazing team of the Polaris project.
The second project that Libre Space Foundation will be mentoring is a “Python Module for RF Collisions”. This project’s goal is to tackle an issue that troubles satellite observers quite frequently. With the number of deployed satellites in constant increase, it is often that satellites transmit with the same or near frequencies. This overlapping of frequencies interferes with the results of the observations and affects their accuracy. Thus, the project we will be mentoring aims at dealing with this exact issue. By building a Python module that will allow the ground station operators to specify the time and the location this interference occurs. This project is closely related and linked to SatNOGS and it will be used by the SatNOGS network as an internal or an external tool to let the observers know which other satellites are expected to be found in the results of their observations.
Google Summer of Code is an annual program offering university students the opportunity to work on open-source projects during their summer break while earning a stipend! Libre Space Foundation is devoted to working on open-source space technologies and you can find out more about our Principles regarding open-source and space in our Manifesto.
This year’s Google Summer of Code application period has been indeed a groundbreaking one as the initiative received 8,902 applications submitted by 6,626 students from 121 countries. These applications were reviewed by 199 mentoring organizations. Eventually, 1,199 students from 66 countries were selected. We are thrilled to be part of this grand initiative. But we are also excited and looking forward to working with our students over the next few months. Congratulations to everyone and welcome aboard!
Pingback: Deep Learning and Space Weather in GSoC’s 2020 Libre Space Foundation contributions – Libre Space Foundation