2023-11-10 23:00:02
A spacecraft will acquire its intellectual abilities in space through artificial intelligence instead of all or a good part of this learning process being carried out on Earth as has been the norm in astronautical applications of artificial intelligence.
After more than two years of development, the SONATE-2 nanosatellite is about to be launched. Launch into orbit by rocket is scheduled for March 2024. The satellite was designed and built by a team led by Hakan Kayal of the Julius Maximilian University of Würzburg, in Bavaria, Germany.
This university has been preparing small satellite missions for about 20 years. SONATE-2 now marks another milestone.
The satellite will test novel artificial intelligence hardware and software technologies in space, albeit close to Earth. The goal is that this type of artificial intelligence can be used in the future to automatically detect anomalies in stars such as planets or asteroids.
This spacecraft will be the first in which artificial intelligence will be trained entirely inside the ship when it is already in space. Typically, this training is done on Earth with powerful computers. But this strategy has weaknesses:
Kayal gives an example of the latter: “Suppose that a small spacecraft must investigate a new asteroid in the solar system in the future. It cannot be trained for this task on Earth, because the object to be investigated is largely unknown. There is no usable data for training, so this must be done on the fly.
Artistic recreation of the SONATE-2 satellite in space, incubating its incipient artificial intelligence. (Image: Hakan Kayal / University of Würzburg)
Sending the new data to Earth first and then training the artificial intelligence by remote control would take a long time for missions to stars far from Earth. A higher level of autonomy based on on-board artificial intelligence would give better results. It would allow interesting objects and phenomena to be detected on the star much more quickly. (Source: NCYT from Amazings)
#Artificial #intelligence #forged #Earth