AI in the Sky: How Artificial Intelligence makes for better Air Traffic Management

At a recent aviation industry event, expert analysts revealed that less than 10% of the data produced by aviation operations in Europe is actually utilised in the running of said operations. That means less than a tenth of the available information is being fully assessed and acted on, leaving more than 90% of that data’s unrealised potential in terms of making commercial flying safer, swifter and more efficient.

Artificial Intelligence is poised to redress this imbalance, with leading industry figures experimenting heavily with this nascent technology with its expansive potential to change aviation as we know it forever. Right now, dozens of AI initiatives and trial runs are underway across the world’s biggest airports, with a view to vastly improve their Air Traffic Management (ATM) capabilities in the face of rising consumer demand.

 

Tackling uncertainty

ATM is all about making sense of uncertain circumstances in order to make the right call every time. With each flight potentially representing hundreds of lives, the margin for error is minute and the consequences of a wrong call are all too often catastrophic. 

While freak accidents can and do occur, a significant proportion of aviation-related disasters could have been avoided if everyone involved in the decision-making process had access to the whole picture. This is the context under which AI is currently being trialled – to give ATM operators better insight into changeable conditions such as the status of flight equipment, ground conditions and, crucially, weather. 

Through the use of ‘neural networks’, AI platforms can quickly absorb and learn from all available datasets offered up by the airport’s operational systems, allowing it to discern what ‘normal’ operating conditions look like. Then, with this target in mind, the system can hunt for any anomaly, no matter how small, and alert human operators and/or take corrective measures itself. This setup allows for all of the airport’s collected data to be fully utilised to give an empowered ATM staff the ‘full picture’ at any given moment, cutting out uncertainty and, hence, cutting down on bad calls from the tower.

 

Roadblocks to Adoption

Integrating AI into any airport’s ATM setup is unlikely to be a trivial task. Traditionally, airports have struggled to avoid data siloes and update legacy systems at the best of times, due to the hugely rigorous demands of the 24/7/365 industry conditions they labour under. For major international airports that handle tens of millions of passengers a year, the thought of a large-scale ICT switch-up is one that brings as much trepidation as excitement. Accordingly, any ATM operator looking to bring AI into the mix will need to create a holistic adoption plan that allows for minimal disruption during the implementation phase as well as during all future operations. 

The recent disaster of the Ethiopian Airlines Boeing 737 crash is a sobering example of the dangers of poor AI adoption and integration, as the failure of the Manoeuvring Characteristics Augmentation System (MCAS) was assigned as one of the key reasons behind the tragedy.

 

Potential Payoff

Augmenting human ATM staff capabilities: Bringing AI into the ATM operational sphere isn’t about cutting out the humans running the system, it’s about augmenting their capabilities with the kind of data analysis and decision-making aids that they simply couldn’t hope to achieve otherwise. 

Reduction in flight delays through greater operational efficiency: The uncertainties caused by weather and mechanical failure can quickly equate to mounting delays as sub-optimal workarounds keep flights grounded or in the air longer than they need to be. 

AI is already proving to be an invaluable asset in this area, with early trials at Heathrow by NATS ATM service showing that its adoption could help to reclaim 20% of lost runway capacity caused by low cloud and reduced visibility. Instead of relying on radar during low visibility to know if flights have left the runway (a process which requires extra safety time built in), NATS is deploying 20 ultra high-definition cameras across the airfield, which feed into an AI platform called Aimee. Aimee tracks all leaving aircraft and informs the tower controllers once they are successfully clear, cutting out the need for extra safety time and potentially allowing Heathrow to reclaim all of its lost 20% runway capacity. 

Boosted safety levels: If it is integrated correctly, AI may quickly come to be seen as an invaluable safety asset in ATM operations, as it will be able to create another layer of checks and anomaly detection protocols that can spot potentially life-threatening circumstances and stop them from ever occurring. This kind of life-saving analysis may extend to predictive maintenance of aircraft and ground equipment, adverse weather condition pattern analysis and more.

 

There’s a long way to go, but it’s down an exciting path

AI adoption in ATM seems all but a forgone conclusion, as the benefits of marrying artificial and human intelligence look like they are too good to miss. However, expect to see a lengthy trialling phase in most leading airports, that will be likely measured in years rather than months, before we see anything comprehensive in their towers. Gradual, considered and intelligent adoption is what’s needed to ensure that AI is a net-plus solution from the very moment that ATM operators use it in earnest.

 

This article was created in association with Airport Show taking place in Dubai on 22-24 June 2020. 

 

Register now for FREE entry to Airport Show, Airport Security and ATC Forum exhibition