Orbiting companies

Whether launching microsatellites – cubesats or building instrumentation that will be used in orbit, space-related research is an increasingly important sector of our economy. And one that will increasingly impact life on Earth.

This is what Elena Donini, a researcher with the Remote Sensing for Digital Earth (RSDE) Unit at Fondazione Bruno Kessler along with Gino Bucciol of Officina Stellare SpA and Mattia Barbarossa of Sidereus Space Dynamics spoke about in the talk “Orbiting Companies” during the ongoing Wired Next Fest in Rovereto.

“In my work, artificial intelligence has allowed us to make some incredible discoveries.  It used to take weeks, even months, to visually interpret a piece of data: now for example we can analyze the entire subsurface of Mars in one afternoon. The use of artificial intelligence and machine learning algorithms makes it possible to carry out certain processes with a speed that was once impossible,”Elena Donini responded to Wired journalist Stefano Priolo‘s question about what horizons the relentless improvement of artificial intelligence in space research opens up.

There are two space missions in which Elena Donini and the research team of which she is a part in collaboration also with the University of Trento are participating: destination Venus with the EnVision mission, while Juice aims to explore Jupiter and its icy moons. For these two missions, the team is in charge of the design, i.e. the design of the radar, and the analysis of the data acquired by the radar itself. These are special radars that allow us to see the subsurface and through which it is thus possible to discover structures or geological processes that otherwise could not be seen.

The focus of Elena’s work at the moment is the definition of artificial intelligence-based algorithms that allow us to extract this information. ized way.  Another relevant aspect of her work concerns “analog data”: since data from these missions will be available years from now – for example, JUICE was launched last year and it will take about eight years to get to Jupiter. Currently, data acquired on Earth are being used to study, for example, the impact of climate change in Antarctica or Greenland, and data from Mars and the moon are also being analyzed to identify cavities and dust accumulations.

For missions that are now active, Artificial Intelligence is also beginning to be used to define orbits: through machine learning algorithms, only the areas of interest for studies are defined and the best orbits are determined to reach them in an optimized way.

“In the future I expect artificial intelligence to be used more and more in all processes in the space domain.  If I have to think of something that seems impossible to me now but that artificial intelligence could improve, I would provocatively point to the use of AI for achieving a launch so powerful that it could shorten the time of space missions from many years to a few weeks,”Elena Donini told the audience.