Airport 2030: How AI will Change Airports and Flying

Staff shortages, climate change or digitalization – aviation is struggling with a wide range of problems.

Whether it’s robot “Ray” at Düsseldorf Airport parking cars by himself in the parking garage, the digital carpet at Geneva Airport making shoe security checks more efficient, or passengers being confronted with apps and automated check-in terminals – the industry’s will to digitize and automate is there, but reality still signals a lot of catching up to do. Unfortunately, the intelligence in devices is still lacking, due to privacy- and security-sensitive regulations and a deep innovation slumber. The industry’s woes are now being addressed more quickly.

Embedded AI as an opportunity

Anyone who thinks of ChatGPT when they think of artificial intelligence only has one part of AI-based support and automation systems in mind. Large language models can, of course, be used to serve the information needs of people at the airport more easily and efficiently. But in many cases, especially where machines are used, AI needs to be even faster and more specialized at the same time. Server- and cloud-based AI like ChatGPT is often too slow and too computationally intensive to solve the “problems of machines.”

That’s where embedded AI comes in – the locally self-sufficient AI variety that doesn’t require network connectivity and can respond in real time. It sidesteps privacy concerns and enables insights, analysis and appropriate consequences to be drawn where previously it seemed impossible. Because embedded AI processes enormous amounts of data in milliseconds, it looks particularly deep and finds patterns and correlations that were previously unknown. In addition, the megatrend is toward decentralized AI, i.e. the next technology level: everything will then be intelligent in its own right, without central networking.

A flight to 2030

Imagine: You enter the airport, counting people at each lock, door, and automated calculation of queues at counters and accumulations in front of security gates is done live by vision and lidar sensors (so-called time-of-flight sensors), providing intelligent operations, real-time counter control and predictions.

You don’t have to fear surveillance in the process, because the embedded AI processes data in the sensor itself, without connecting or sharing the data with the network, Google and co. In addition, the response is immediate and independent. Thus, even the interaction with automatic check-in or drop-off counters can be equipped with better user interaction through gestures, speech and object recognition with luggage privacy. The acute and chronic shortage of personnel has thus been helped considerably, and the hygiene lessons of the pandemic have certainly also been learned.

Even more security and faster processes

You will also encounter embedded AI on your way through the security gates, where intelligent NIR sensors (NIR: near infrared) can detect stressed persons with security risks, anomaly detections in the body scanner or in the XRAY CT detector work much more profoundly than today’s personnel checks – for even more security and faster processes.

And as your suitcase makes its way through the underground basement mazes of the transportation and logistics system at any major airport, all that matters is safety, time and reliability. Here, embedded AI with appropriate vision, radar and lidar sensors will provide more precise and faster sorting and control. In addition, laser, ultrasonic or vibration sensors will monitor weighing as well as each drive and roller of the conveyor belt for wear and tear and failure prediction (predictive maintenance), because failure is expensive and unplanned is very personnel-intensive.

We jump to the tarmac: here, vehicles for pushback or logistics transport are also good examples of failure prediction, early failure detection or automated obstacle detection by AI in the vehicle. Incidentally, this also applies to the supply units for heating or air supply.

Predictive maintenance makes actuators safer

Last but not least, we look at the opportunities for local AI in the aircraft itself, once you’re sitting inside: Pressure and vibration sensors can monitor (injector) combustion in turbines in milliseconds – also for the purpose of kerosene mixture analysis – and immediately adjust or optimize it for fuel efficiency. Every actuator component in the aircraft can also be made safer with predictive maintenance, making maintenance cycles much more flexible as regulation catches up. User interaction with pilots or in the passenger cabin for comfort can also be rethought. Even flight maneuvers or control algorithms can benefit directly from AI on board.

But who says there won’t be electric air cabs by 2030? Again, propulsion monitoring for the purpose of fail-safety for the purpose of passenger adaptation of flight technology is critical, as are autonomous support functions that AI will inevitably require to run locally on board. Embedded AI, then, will inevitably lead the aero industry into the future, and it will do so in a technology-open way.

Viacheslav Gromov

Viacheslav Gromov

is Managing Director of AI specialist AITAD.