A satellite the size of a cereal box, carrying a camera and an AI chip, is now in orbit.

Intel (NASDAQ: INTC) has the distinction of launching the first onboard AI processing chip into space. Earlier this month, the European Space Agency and Intel announced the successful deployment in space of PhiSat-1, the first-ever satellite with onboard AI-processing capabilities. (Business Insider)

Launched from a rocket dispenser on September 2, the PhiSat-1 is positioned about 530 km above our heads. It is moving at a speed of 27,500 km per hour in a sun-synchronous orbit.

PhiSat-1

The satellite’s objective is to monitor polar ice and soil moisture, as well as to test inter-satellite communication systems.

The satellite carries a hyperspectral-thermal camera and an Intel Movidius™ Myriad™ 2 Vision Processing Unit (VPU). The latter is responsible for the AI heavy lifting operations onboard the spacecraft.

Myriad’s immediate function is to curate the huge mass of data captured by the camera.

AI at the ultimate edge – space

The big problem facing the scientists was the sheer volume of data generated by the hi-fidelity camera onboard the PhiSat-1. The camera unfortunately does not know how to differentiate between a cloudy and clear environment.

It, therefore, takes a large number of photographs that are useless because, at any given time, clouds envelop two-thirds of the earth’s surface.

The junk photos consume precious internet bandwidth to send down to earth. After all that, scientists would likely delete the unclear photos.

The scientists decided to use onboard AI (also known as “on edge” processing) to curate the photos. Myriad-2 would examine the images, trash the useless ones, and send only the good ones to earth.

By discarding the cloudy images at the source, they saved nearly 30% of bandwidth.

“Artificial intelligence at the edge came to rescue us, the cavalry in the Western movie,” says Gianluca Furano, data systems and onboard computing lead at the European Space Agency.

“Space is the ultimate edge,” says Aubrey Dunne, chief technology officer of Ubotica, the Irish startup that built and tested PhiSat-1’s AI technology. “The Myriad was absolutely designed from the ground up to have an impressive compute capability but in a very low power envelope, and that really suits space applications.”

Ubotica worked with cosine, the maker of the camera, in addition to the University of Pisa and Sinergise.

After three weeks of testing, the team could establish that Intel’s Myriad AI onboard the PhiSat-1 was working fine.

ESA then announced “the first-ever hardware-accelerated AI inference of Earth observation images on an in-orbit satellite.”

Satellite-as-a-service!

Scientists can now visualize multiple applications of AI on satellites.

For example, the satellite, during one orbit, could switch from spotting wildfires on land to rogue ships or environmental accidents at sea such as oil spills.

It could measure crops and soil moisture over farms and forests, and assess the ill effect of climate change on melting ice caps.

Related Story:   Satellites and AI Could Together Predict Wildfires Accurately