02/05/2023
1035 views
10 pleases
Despite the rain dripping from the windows, the atmosphere in this room is bright. Twelve teams have gathered in ESA’s ESTEC technology center to explore how we can use the latest developments in artificial intelligence and advanced computing to make satellites smarter. To make them more responsive, agile and independent.
The 17 scientists and engineers who attended the meeting in person, including three colleagues from ESA. Three others joined online.
Smart speakers. Autonomous cars. Smart heating. We are seeing these inventions making their way into our daily lives and terms like ‘the cloud’, ‘blockchain technology’ and ‘machine learning’ are becoming more familiar. What if satellites could use existing technologies that collect data, make sense of it, and act accordingly without human intervention to become smarter?
By nature, satellites take a long time to develop, while the digital revolution moves very fast on Earth. Compared to the processors we use on Earth today, computers aboard satellites are often outdated. Additionally, the space industry has a history of being conservative, taking only necessary risks and relying on space-qualified hardware and computer techniques wherever possible.
The 12 teams have all spent six months exploring how to accelerate the adoption of the latest AI capabilities in space and are meeting to discuss their findings. Some have investigated how smarter satellites could directly improve our lives on Earth, such as by better detecting methane leaks and managing disasters from space. Others have looked into how smarter satellites could support more sustainable exploration of the Moon and make lunar rovers more independent.
“Many satellite applications require a rapid response, for example in disaster management or tracking moving objects,” says Gabriele Meoni, an ESA researcher who led the meeting. “If satellites could process data directly and independently in space, we could really improve the way we handle these situations.”
The FSSCat mission to two CubeSats, Europe’s first artificially intelligent Earth observation mission
Gabriele continues: “I was really surprised by the results of this suite of research studies. Some have shown that using AI, results can be obtained that are comparable to current methods but much faster”.
Historically, satellites have been designed to collect a lot of data and send it to Earth, where engineers process the data to extract useful information. Gabriele adds: “I want to clarify that bringing artificial intelligence on board satellites to process data is not a replacement for the current system, but rather a complement to it, as it allows users to get a faster view of the data. I think that with this research we have taken a good first step towards understanding how we can do it better.”
Most teams found that the benefits of smarter satellites go beyond sustainability and could lead to financial opportunities, helping to build a strong commercial space ecosystem in Europe. During a round table, the teams discussed the commercial potential of this technology, highlighting the market value of some products.
A diagram describing the work of one of the projects, which explored the use of a peak and peak system to monitor methane emissions. An undedicated satellite (the tip) rapidly scans the Earth; when it locates a suspected methane source, it tasks another spacecraft (the cue) equipped with a high-resolution detector to observe it in more detail.
“In the recent ‘Revolution Space’ report, artificial intelligence and cloud computing were identified as examples of where Europe has historically lost out on potential markets,” explains Leopold Summerer, head of Advanced Concepts at ESA. “This series of activities marks a change, helping European companies to grow and thrive in an innovation-driven global ecosystem.”
ESA Discovery funding offers established and new space and non-space companies the opportunity to explore emerging fields even when they are not yet profitable. It allows them to experiment and take risks. It means that European companies can not only be ready for future space markets, but play a vital role in shaping them.
A diagram describing another project that explored methane leaks at gas plants in near real time. Earth observation and weather satellites would be combined with artificial intelligence to detect and measure leaks.
“The results of the activities suggest that smarter satellites could bring huge benefits, but this is only the first step,” adds Gabriele. “For me, the next step is to increase the technology readiness levels of the various concepts, working on early prototypes for ideas that propose new technologies and filling the technology gaps for those that show business potential.”
It seems appropriate that the final project presented during the meeting was called ‘FRIENDS’. At the end of the event, attendees shared business cards and “see you soon.” It will be exciting to see how the collaborations and competitions initiated through these ESA-funded activities propel the field of cognitive cloud computing forward in space.
Depictions of 11 of the projects
The call for ideas on ‘cognitive cloud computing in space’ was conducted through the Open Space Innovation Platform (OSIP) by ESA’s Foundation Activities Discovery element together with the Agency’s Laboratory and Commercialization Department. The activities were funded from the Discovery element using the ESA Initial Support for Innovation (EISI) programme; each ran for six months with a budget of EUR 100,000.
Read descriptions of all activities here.
#ESA #extends #artificial #intelligence #cloud #computing #space