Intelligent steps for the future

An abstract representation of AI in the palm of your hand.

Article by
Tanja Grobotek
European Affairs Director


Artificial intelligence (AI) – and its machine learning applications in particular – has captured the attention of many leaders and researchers worldwide for its potential to enhance the performance of different industries, including the aviation sector.

It is an area of computer science that includes highly advanced computational methods that mimic the way the human brain works, and is a powerful technology that can find patterns in massive unstructured data sets and improve its own performance as more data becomes available.

Today, typical AI capabilities include speech, image and video recognition, smart automation, advanced simulation, as well as complex analytics and predictions.

The application of AI in the aviation field can help to overcome the current challenges of capacity, environment, connectivity and security.

As highlighted in CANSO Europe Vision 2035, CANSO believes it is paramount to find a clear journey towards system-centric technology, increased automation, oriented architecture, virtualisation and interoperability to transform air traffic management (ATM) performance. Digital transformation of the aviation industry enables data-driven technologies like AI which form a critical part of the highperformance ATM systems of the future.

CANSO is focused on building a stronger future today – and some member air navigation service providers (ANSPs) are already actively harnessing the power of AI to improve current processes, and create new ATM applications.

For example, DSNA in France is developing new AI tools to advance operational solutions to deploy the optimal configuration of sectors, and thus to optimise capacity with available resources. European airspace is divided into several elementary blocks of airspace which allow modularity and flexibility in building the different airspace configurations to meet expected and effective traffic flows. In the case of France, its airspace is composed of 168 elementary sectors.

In the tactical phase, Supervisors and Flow Management Positions (FMPs) analyse, on a continuous basis, traffic demand and available capacity from the analysis of the operational air traffic control centre (ACC) data, including resources, technical availability, unexpected events and environmental data (weather, military activity). With their expertise, they manually determine the most adequate sector configurations. Without automation, identifying the sequence of adequate ACC sector configurations according to traffic flows and associated workload is time-consuming. And only solutions from predefined catalogues are routinely used, limiting the possibilities.

DSNA is therefore using innovative techniques, offering dynamicity, optimisation and time-saving in capacity management and ACC sector configuration. In order to maximise the benefits of airspace modularity while optimising resources, automated functions have been developed to support Supervisor and FMP decision-making. A new AI tool is able to optimise airspace solutions in case of high workload and unforeseen events which could require quick decisions for airspace re-organisation. It also provides additional means in the Air Traffic Flow Control Management (ATFCM) toolkit using dynamic sector configurations to facilitate users' preferred routings.

In the UK, NATS is using AI to measure air traffic characteristics (for example, high traffic volume, weather conditions) and predict the likelihood of potential safety events, like runway exit points for example. This will enable strategic corrective actions to avoid them. NATS is also exploring how AI could be used to help reduce flight delays. A project is now underway, within NATS’ bespoke Digital Tower Laboratory at Heathrow Airport, to test whether a combination of ultra HD 4K cameras along with state-of-the-art AI and machine-learning technology can be used to help improve the airport’s landing capacity in times of low visibility and improve punctuality (see NATS article in this issue). Non-operational trials are now underway to understand the feasibility of introducing the technology into service as early as this year.

The safe and fair integration of drones, or unmanned air systems (UAS), into the aviation system will be a major change project not only for ANSPs but also for the industry as a whole. The evolution of ATM for commercial drone systems relies on greater digitalisation, automation and pace compared to conventional air traffic management evolution.

AI is seen by many UAS operators and manufacturers as an essential prerequisite for communication between drones. In the event of irregularities, they must be able to keep drones on track and react to changes in milliseconds, and adapt their own flight routes according to the information provided by other drones.

Interested? Subscribe!

Never miss out on the latest developments from EUROCONTROL

In the longer term, one of the most promising benefits of AI is reducing human stress by supporting controller decisionmaking, which will improve operational capability, the consistency of delivery and will enhance safety.

Other AI applications being explored by ANSPs include enhanced ATCO training and simulations, improved predictive maintenance of air traffic management/communications, navigation and surveillance (ATM/CNS) systems and better trajectory prediction.

There are many other tasks that could be performed more efficiently by machines in advanced ATM systems, supported by big data and AI rather than human operators. However, there are several key aspects that need to be addressed first.

It is difficult to assess, validate and certify the AI systems, especially if they evolve in time, in particular for low occurrence (for example, safety) events. And for non-safety critical operations, certification needs to be less stringent than for safety-critical operations. This means that regulationsneed to be performance-based, and based on the nature of each AI solution. 

Going forward, as ATM systems modernise and the use of advanced technologies increases, it will be essential to assess the impact on human operators (controllers, pilots, engineers and so on) and organisational aspects. ANSPs will therefore need to assess the role of the human in the future ATM system, training needs, user acceptance and regulatory changes (liability).

Finally, as it was reported at EUROCONTROL’s recent Artificial Intelligence in Aviation event, less than 10% of the data produced by the industry in Europe is actually used. As machine-learning algorithms that make predictions have to be trained using historical datasets, it is essential to define clear data governance to ensure availability and integrity of data in order to optimise its use for more efficient and effective air transport.

For its part, CANSO is committed to working closely with all stakeholders and regulators to address the above issues, sharing best practice and ensuring that the ATM industry is fit for the future.

CANSO has recently launched a strategic technology initiative, identifying and assessing emerging technologies for potential impact and benefit for ATM. This will provide an important venue for collaboration and information among CANSO’s global ATM network of air navigation services providers and suppliers to the industry, allowing key players to share information on new ATM technology and other related technologies and encouraging adoption and implementation.

The ATM industry has always been at the forefront of safe and efficient innovation and AI holds huge potential to develop and enhance the industry’s capabilities, from increasing capacity and optimising traffic flow, to safely integrating a diverse range of airspace users. The challenge ahead will be how the industry can embrace the technologies available, learn from each other and unlock its full potential.

Read the full edition of Skyway Autumn/Winter 2019

About the Civil air navigation services organisation (CANSO)