Tech UPTechnologyThe first satellite with AI on board is already...

The first satellite with AI on board is already in orbit

On September 2, an experimental satellite the size of a cereal box was launched from a rocket dispenser, along with 45 other satellites of similar size. The first satellite with artificial intelligence (AI) to be launched into space, called PhiSat-1, now rises at more than 27,500 km / h in a synchronous orbit to the Sun, at an altitude of about 530 km.

The PhiSat-1 features a thermal and hyperspectral camera, as well as on-board AI processing thanks to the Intel® Movidius ™ Myriad ™ 2 Vision Processing Unit (VPU), the same chip as They have many smart cameras inside and even a $ 99 drone. PhiSat-1 is, in fact, one of the satellites on a mission to monitor polar ice and soil moisture, while testing inter-satellite communication systems to create a future federated satellite network. .

Problem to solve

The first problem Myriad 2 is helping to solve is how to handle the massive amount of data generated by high-fidelity cameras like the one on the PhiSat-1. “The ability of sensors to produce data increases by 100 with each generation; however, our ability to download them increases, but only by three, four or five in each generation, ”says Gianluca Furano, data systems and onboard computing lead at the European Space Agency (ESA), who led the project. collaboration on PhiSat-1.

About two-thirds of the surface of our planet is continuously covered in clouds. For this reason, usually, many images of clouds that do not work are captured and that are saved, sent to Earth through a significant download bandwidth, saved again and reviewed by a scientist or an algorithm on a computer for hours or days later, so that, finally, they have to be eliminated. “And the AI came to rescue us, like the cavalry in a Western movie,” says Furano. The idea the team came up with was to use on-board processing to identify and discard cloud images, thus saving around 30% of bandwidth.

Computing capacity

“Space is the cutting edge,” says Aubrey Dunne, Chief Technology Officer at Ubotica, an Irish startup that built and tested the AI in the PhiSat-1’s in collaboration with Cosine, the camera’s manufacturer, the University of Pisa and Sinergise to develop the solution. complete. “Myriad was completely designed from the ground up to have impressive computing power, but at very low power. This is fully tailored to the needs of space applications. “

However, Myriad 2 was not originally intended for orbit. Typically, spacecraft computers use highly specialized “radiation-hardened” chips that can be “up to two decades behind leading-edge commercial technology,” Dunne explains; and, at the time, AI was not part of the equation. For this reason, Dunne and the Ubotica team performed the “radiation characterization,” putting the Myriad chip through a series of tests to figure out how to handle any errors or deterioration that might arise.

Chip with radiation

ESA had never tested a chip with this radiation complexity before, and the project promoters doubted whether they would be able to perform their tests correctly, to the point of having to develop the manual on how to perform a comprehensive test and characterization from scratch. of this chip. The first test, which was 36 hours of radiation in a row at CERN at the end of 2018, “was a very stressful situation,” Dunne says. However, that test and his two follow-ups “luckily we went well.” Finally, Myriad 2 passed the test in its standard form, with no modifications required.

After successful testing, this low-power, high-performance computer vision chip was ready to venture beyond Earth’s atmosphere. However, another challenge came. Typically, AI algorithms are built or “trained” using large amounts of data to “learn”, in this case, what is a cloud and what is not. But since the camera was so new, no data was available, so an application had to be prepared with synthetic data extracted from the existing missions.

More than a year of waiting

All this integration, and system and software testing, involving half a dozen different organizations from across Europe, took four months to complete. Unfortunately, a series of outside events – rocket delays, the coronavirus pandemic, and unfavorable summer winds – caused teams to wait more than a year to find out if PhiSat-1 would function in orbit as expected.

The September 2 launch from French Guiana was fast and flawless. For initial verification, the satellite saved all the captured images and recorded each of its artificial intelligence decisions regarding cloud detection, so that the team on the ground could verify that its implanted brain was behaving as expected. After three weeks of uncertainty, Pastena was able to proclaim that they had become part of the “history of space”. ESA announced that the team was “happy to be able to present for the first time images of the Earth from an orbiting satellite obtained thanks to the first inference of hardware-accelerated AI.”

Useful pixels

By just sending useful pixels, the satellite will improve bandwidth utilization and significantly reduce additional downstream costs, not to mention saving time for scientists on the ground. In the future, the uses for teensy low-cost AI-enhanced satellites will be countless, especially if the ability to run multiple applications on them is added.

When flying over wildfire-prone areas, a satellite is able to detect fires and notify local response teams in minutes instead of hours. In the case of the oceans, which are normally ignored, a satellite can detect illegal ships or environmental accidents. In forests and farms, a satellite is able to detect soil moisture and crop growth. In ice, you can track thickness and thaw to help monitor climate change. Many of these possibilities will be tested very soon: ESA and Ubotica are working together on PhiSat-2, which will bring another Myriad 2 into orbit. PhiSat-2 will be “capable of running artificial intelligence applications that can be developed, easily installed. , validated and operated from the spacecraft during its flight using a simple user interface ”.

Chip war rages on: TSMC suspends Biren chip manufacturing

The semiconductor company is seeking to follow US rules, which prevent companies from developing certain technologies.

Do you want your company to be more optimal? use digital twins

According to one study, organizations that use them have seen a 15% improvement in operational metrics.

The brain works like a quantum computer

New research from Trinity College Dublin concludes that certain brain functions 'must be quantum'.

Not everything is DALL-E: The best websites to create images with Artificial Intelligence

The use of Artificial Intelligence in creative work is becoming more and more common and these are some of the platforms that you can use.

This is Optimus, Tesla's humanoid robot

Impressed? Tesla founder Elon Musk wants to build millions of robots like these and sell them for 20,000 euros a unit.