A featured contribution from Leadership Perspectives: a curated forum reserved for leaders nominated by our subscribers and vetted by the Construction Tech Review Advisory Board.
Swissport
Cognitive Computing at Swissport
Global aviation has been growing at an annual rate of 45 percent. At the same time, the capacity at airports, especially in western Europe, is limited. The high density off lights adds complexity to air traffic control and adds to delays. For airport ground service providers, this has a major impact on shift planning and task assignment. Arising share of non-planned short connections, caused by delayed incoming flights, adds additional complexity to effective planning and task assignment.
Generally speaking, a lot of information is available. The trick is to have it available in aggregated form, at the right time—ideally real-time—and at the right place. While schedule planning and aircraft tail-assignment are relying on optimisation algorithms, short-term task assignmentis still done mainly by human operators on complex user interfaces. Moreover, this often happens on very short-notice and is reactive by nature. Humans can manage a lot based on experience, but they can only take into consideration a limited number of factors. For task assignment, it would be helpful to use information such as location of gate, aircraft type, team expertise, number of transit passengers, priority baggage, available ground handling vehicles and positions, time for passengers to move from one location to another, forecasts on queuing times at security, relying on real-time information and more.
In the case of Helsinki, Swissport can predict the expected arrival times of flights during the day, based on continuous data collection and algorithms, which were trained with several years of data. The system can highlight when a delay of five minutes of a certain flight in the morning will result in a delay of half an hour or more during the day, as the delayed aircraft flies to different destinations later in the day. The resulting delays are influenced by the traffic situation, both on the airways and at the airports where the aircraft is scheduled to pass through or arrive during the day.
Based on such delay predictions and the continuous learning of algorithms, resource planning can be optimised and refined. Sharing the resulting insights with the airport and airline helps them optimise schedules and capacities and prioritiseflights and assign gate positions.
While humans are focusing on “task tetris” and move lots to teams, ongoing machine learning algorithms can take into consideration multiple data sources and make scenario calculations and compare them. This enables us to predict future actions and assign tasks more effectively.
Cognitive computing allows airport ground service providers to reprioritise on-the-go and to be more flexible in their co-operation with airlines. In case of a delayed incoming flight with a lot of transit passengers to international destinations, it might be better, for example, to prioritise this delayed flight outside of the normal service-level-agreement and to take resources off an on-time regional flight with just a few passengers. Instead of having 100 long-haul passengers stranded at the airline’s hub, the airline can decide to focus on this high value flight first. This supports a better customer relationship and opens new revenue streams through value generation for the client. We can improve service efficiency and simultaneously contribute to enhancing the passenger journey.
Swissport is extending the use of machine learning and artificial intelligence to more and more of its locations around the globe. Operating at more than 300 airports world-wide, the company is ideally positioned to build on its vast source of quality data and help airlines deliver a better passenger and customer service. This also opens perspectives regarding offering this “data centric approach” for partners and contributes to the development of new services with and for employees and customers.
