Can AI organise our skies?

While the skies over our heads have been quieter throughout the pandemic, they’re normally jammed full of planes: Heathrow alone saw 1,300 flights take off or land each day in 2018, and ferries more than 80 million passengers annually. That makes air traffic control a complex job — and with unmanned drones and small electric planes eventually added to the mix, it could become completely unmanageable.

That’s why NATS — once known as National Air Traffic Services — has teamed up with data scientists and AI researchers at the Alan Turing Institute, the national body for investigating and developing such smart technologies.

Since 2018, the two organisations have been working on project funded by the Engineering Physical Sciences Research Council worth £12 million. Called Project Bluebird, a sweet-sounding name for a serious bit of work, it has the high-flying aim of bringing automation and AI to air traffic control.

▪ The legacy effect

The first stage of the project involved developing a proof of concept, says Dr Evelina Gabasova, principal research data scientist at Turing, and that revealed an early hurdle: AI and air traffic control are two disparate fields. “As you can imagine, we speak different languages in AI research compared to air traffic control,” she said. “So