AI in the Sky

Artificial Intelligence (AI) holds a great deal of promise to assist organisations with enhanced information to support tasks that still need to be led by a human. Air Traffic Management (ATM) is one of those areas where we’ve only seen baby steps to date due to the safety-critical nature of these tasks. Keeping planes safe in the air and on the ground will likely always have a human involved. That’s not to say there aren’t opportunities to aid the individuals performing these tasks.

Nokia has an AI-based solution that just does this. Today, siloed systems in an operations centre have to be correlated by a human in order to determine the full situation awareness.  An integrated command and control centre concept is already deployed in smart cities, but could be translated into a solution for air navigation service providers (ANSPs).  A unified user interface – single pane of glass – tracks information provided by IoT sensors and video in order to have a single view of the situation at the airport.

Let’s consider a couple of ANSP related examples.  One possible outcome is to create an ANSP communication, navigation and surveillance (CNS) operations and maintenance module.  Based on visual inspection on a regular timeframe of ANSP assets, a predictive maintenance program through the use of IoT sensors to monitor the conditions of these assets, could warn that approaching maintenance should occur.  Feasibly this could save up to 30% in maintenance cost and time spent by personnel in the field to check assets.  Along with an improvement in availability of assets, reduced ongoing maintenance and increased infrastructure utilisation.

Another possibility is to deploy a smart approach and tower operations solution.  The basic concept is to gather information from various tools and sensors already in existence. This could include weather forecasts, flow management, sensor data – IoT, video, computer aided dispatch (CAD), alert monitoring and CNS. All data would be accessible via a secure data lake in order to allow for historical and predictive modelling.  AI-based data analytics could then help to suggest actions to be taken for predictive runway maintenance, slot allocation and situational awareness.  These are just a few examples of the possibilities.

Below are two visualisations as to how an airport could create a COVID-19 monitoring solution based on accessing passenger information from existing sensors.  The first shows the tracking programs that can compile the actual events and interventions required over the past week.

 

The second shows resulting alerts when any measures outside of the set parameters occur.  This could also be linked to deployment of personnel to address these issues.

Are ANSPs ready to hope on board the wider use of AI?  I think it will still take a while to grow in wider usage.  The implementation of tools such as the Nokia solution will allow ANSPs to introduce AI in areas that are not safety-critical and get comfortable with how this technology could migrate into more areas that directly impact ATM.