We hear an awful lot about Artificial Intelligence (AI) making waves in Healthcare, Pharmaceuticals, Finance and of course retro computer games. Less reported is that Air Travel is also undergoing a revolution, as AI’s disruptive influence begins to take effect in this fast-paced and rapidly growing industry.
McKinsey [Forbes] predicts that AI will have an 11.6% impact on Travel Industry revenues – the largest increase of all industries! What’s more, compared to the more traditional techniques in use today (pre-AI statistical methods), Travel will see that greatest benefit of this AI introduction.
In further support of the above, SITA state that 52% of airlines plan major AI programs by 2021, as well as 52% of airports putting AI firmly on their R&D horizon. Predictive Analytics feature at the top of the AI wish-list.
So, we’re in for some exciting times ahead.
But amongst all this speculation and excitement, it’s often difficult to see any concrete examples of exactly how AI is being used in the Travel Industry today. Here are four examples of real AI projects, funded, completed and delivering real benefits for Air Travel.
No 1 – Predictive Analytics for Immigration
Passenger load at immigration desks is notoriously difficult to predict. Very clever people have been trying to statistically model this for a long time with little or no success. Aurora-AI has recently delivered a new AI that predicts the number of passengers arriving at immigration desks, hours ahead of time, with precision not seen before, with nothing but a flight schedule and passenger numbers to go on. The results are so astonishingly accurate, it almost seems like magic. Any update to aircraft arrivals are immediately taken into account by the AI, providing a complete picture that evolves as operational events unfold. Long delays building up at immigration could soon to be a thing of the past, as rotas and tea-breaks are dynamically adjusted to ensure the operation is prepared for unexpected busy periods.
No 2 – Closing the Gap in Luggage Processing
Bar-coded bag-tags have long been used to process hold luggage and it works reasonably well, most of the time. But that’s not good enough, not when it’s your bag that goes missing. So, Aurora developed an AI that reads flight code text from bag-tags at lightning speed at any angle, curvature, smudged, skewed and upside-down, whilst bags whiz past on the conveyor. This AI is utilising powerful GPUs to apply real-time OCR at an unprecedented level of accuracy. The system is so successful that it achieves almost (1% difference) the same read accuracy as the laser barcode scanner itself. With both systems combined the miss-rate is tiny – plus it continues to operate even when the datalink to the barcode decoding system is down.
No 3 – Self-Boarding
Heathrow recently announced the world’s biggest biometric rollout in the Air industry. The technology behind this momentous endeavour is an Artificial Intelligence that has been evolving at Aurora over the last few years, whittling down the face recognition error rates until they are almost non-existent. With every passenger’s face recognised by an ever-vigilant AI, not only are we safer, with a more secure border, but the journey is faster, with less hassle and more time to relax. Aurora’s new Orca face recognition engine can match faces in infrared (immune to lighting conditions), authenticate against a passport image (chip or scanned page) and verify back against a mobile-acquired selfie (for home registration).
No 4 – Advanced Threat Detection
One of Aurora’s AIs recently featured on BBC news with our partner Sequestim in an advanced body-scan prototype. This new device can detect threats on the body of a passenger, in real time as they stroll past, without even breaking a stride. Aurora developed the AI which analyses the image output form the advanced terahertz scanner to detect and classify threats. No more standing with your arms in the air! The AI employs two Deep Neural Networks: one to detect a threat; and a second AI to classify it as a gun, knife, powder bag, etc.
So that’s my AI list. Just a select few of Aurora-AI’s deployments; all real projects, delivered, proven and revolutionising Air Travel.
So next time you’re traveling through an airport, spare a thought for the AIs working tirelessly on your behalf. They predicted your arrival to ensure desks are appropriately manned, caught your luggage with the dodgy barcode just in time for loading on the plane, screened your fellow passengers for threats and eliminated the queues to expediate your journey with minimal hassle. That’s real–world Artificial Intelligence today.
I wonder what the 52% of airports and airlines have in mind for AI this year? Do get in touch and let me know. If you’re attending PTE 2019 in London ExCel (26th – 28th March), you’ll find us on stand #1090.