ESA and Irish start-up blast space AI chip with intense radiation

29 Nov 2018

The Intel Myriad 2 AI chip for smart image processing being placed in the beam line of the Super Proton Synchrotron particle accelerator at CERN. Image: CERN/Maximilien Brice

To test the latest in spacetech, ESA asked Ireland’s Ubotica Technologies to blast a new, potentially space-bound, AI chip with powerful radiation.

There was a distinct Irish presence in Europe’s progress in spacetech recently following a project led by the European Space Agency (ESA) to test Intel’s new Myriad 2 artificial intelligence (AI) chip, developed with Irish expertise.

The chip uses AI for high-performance, low-power vision processing and can be pre-trained with data to recognise particular features and patterns, or perform in-depth 3D sensing. Its architecture was originally developed by Irish start-up Movidius, which was snapped up by Intel in 2016.

Co-founded by Inspirefest 2017 speaker David Moloney, the company had previously struck a major deal with Google to bring human vision processing and a lot more to future smartphones, drones and robots through one of its chips.

A massive particle accelerator

Now, in an effort to judge its ability to withstand the rigours of space, ESA has recruited another Irish start-up to blast the chip with one of the most energetic radiation beams available on Earth.

Working with Ubotica Technologies headquartered in Dublin, ESA put the Myriad 2 chips in the path of an experimental beam line fed by the Super Proton Synchrotron (SPS) particle accelerator.

Located in a circular tunnel nearly 7km in circumference, the SPS is CERN’s second largest accelerator after the famous Large Hadron Collider, which the SPS feeds into in turn. After donning hard hats, venturing into the ground floor ‘cave’ surrounded by protective concrete and placing the chip in the beam path, the Ubotica and ESA members retreated to a safe distance and fired.

After the successful test, the results will now be studied to see how the chip copes with some of the most extreme radiation it could face when performing in-orbit image processing on future space missions.

Diagram showing intense solar radiation striking a satellite.

Charged particles emitted from the sun, confined within Earth’s magnetosphere or originating from the wider universe, are a major cause of satellite anomalies and malfunctions. Image: SSA

‘It’s a good result for all concerned’

“AI is a way of boosting the performance of any system with a camera in the loop,” explained ESA on-board computer engineer Gianluca Furano.

“By autonomously figuring out the distance of an object from a camera and how fast it is moving, it can take many more and better images. This also offers a means of enhancing guidance, navigation and control – for instance, to capture drifting items of space debris.”

ESA pointed to Movidius’s original purchase of the LEON4 core from Cobham Gaisler in 2016 as being key to the success of Myriad 2.

“It’s a good result for all concerned because with potentially hundreds – already tens – of millions of customers making use of the LEON architecture, we gain a larger user base and feedback to enhance design quality,” Furano said.

Right now, ESA is already looking at how to use Myriad 2 in space, including putting it on Italy’s Tyvak Mark-I cubesat, which will carry the HyperScout hyperspectral imager. It is also being considered for maritime vessel recognition use, based on the on-board integration of ‘automatic identification system’ signals from ships.

Colm Gorey was a senior journalist with Silicon Republic