Making airport surface traffic more efficient

Researchers at SINTEF have been working to optimise road and rail traffic patterns for more than 25 years. More recently they have being applying their models to aircraft traffic movements. It all started with optimal taxiing planning and guidance to and from the runways. The optimization group is now in the middle of its third project.

"Results show that our optimization system is able to identify the optimal timing and taxiing sequences for runway use by a given aircraft," says Patrick Schittekat at SINTEF. "It can plan detailed taxiing routes, recommend new routes automatically, and respond to a continuous stream of updated information. It is now possible to improve the punctuality of aircraft departures by up to 60 per cent," he says.

Holistic insight is key

There are many factors that currently restrict efficient taxiing at modern airports. These include the use of a multitude of software systems and the fact that traffic is being controlled from different places. Traffic controllers supervising aircraft take-offs and landings currently sit separately in their own towers. The systems that coordinate their actions have to be improved.

"Aircraft taxi towards the end of their take-off runway and wait in line before they can depart," says Schittekat. "Waiting uses up a lot of fuel, and in cases where many planes are scheduled to take off at the same time, waiting times of 5 and 10 minutes are not unusual. There is potential to achieve 30 per cent reductions in waiting times, and improve safety into the bargain," he says.

The optimization model takes into account the basic rules and physical restrictions that apply at all airports, such as safety zones around turning aircraft and one-way taxiing routes.

The model is designed to set up routes that avoid taxiing conflicts. It also defines a safety zone around each plane that varies in size and shape according to factors such as aircraft speed, wingspan, and the nature of its present activity. Optimal surface traffic movements require that these zones must never overlap.

Optimization

Researchers are attempting to improve the control tower software system using mathematics and optimization techniques. Traffic controllers receive a constant stream of data, and they need to be updated at all times on the traffic patterns that make best use of runway capacity.

The optimization model is being continuously supplied with data on the status of all aircraft movements on the ground and in the airspace surrounding the airport. This information is sourced from two other systems supplied by the Swedish and French companies SAAB and Thales, respectively.

"Our model then responds by sending recommended landing and take-off times back to these two systems before receiving more data. The system works as a perpetual loop," says Schittekat. "For different reasons, some aircraft will not be able to adhere to a given plan, but new information allows the algorithms to make modifications so that a new recommendation can be sent -- all in the space of 0.05 seconds," he says.

Researchers are aiming to identify the best way of delegating work between man and machine. "We need more research to identify the ideal means of coordinating work between our optimization system and the traffic controllers," says Schittekat. "This is why we've launched an internal strategic project just for this purpose," he says. (http://www.sintef.no/nextgendst).

Fewer aircraft taxiing around the airport

The Norwegian researchers have been working closely with Swedish Air Traffic Management Services (LFV) and have carried out comprehensive tests and simulations using Arlanda airport in Stockholm as a test case. LFV's extensive simulation platform located in Malmø is almost like the real-life Arlanda. There is a room dedicated as a take-off tower, a second for landings, and a third controlling taxiing routes. There are also about five or six pilots present.

"Arlanda has tested the take-off/landing combination before, but without much success," says Schittekat. "They were testing a fixed, repeatable, pattern such as two take-offs and two landings during a two-hour period. But real traffic patterns are not sufficiently regular or predictable to make this approach effective. Take-offs and landings have to be merged to a much greater degree as part of a dynamic system. We tested the different systems for an entire week and observed how the air traffic controllers respond and communicate when take-offs and landings are combined. It emerged that the controllers experienced lower levels of stress using the merged system," he says.

Simulations and tests have also been carried out at Hamburg airport. Here too the results showed that the new system promoted better traffic flow. There were fewer aircraft moving around the airport, and the controllers experienced lower levels of stress.

Award

The project "Multiple airport arrival/departure management" is now concluded. The software system developed by the researchers is now ready, and can communicate with the other systems.

The project recently received an "Outstanding Project" award from the EU's SESAR Joint Undertaking programme. The SINTEF researchers participated in a small, but important, part of the project, involving combining technologies from Saab and Thales that has since been validated by the Swedish LFV. In connection with the award, the project was cited on the basis of its level of innovation. SINTEF's work has borne fruit and will be continued as part of several projects making up the major EU aviation research programme SESAR 2020.

"SINTEF has a major advantage in that we are able to combine technologies and expertise within the organisation," says Schittekat. "Few people have associated us in the past with expertise in air traffic control, but in the last three to four years SINTEF has consolidated its skills and can now 'talk the talk' with the rest of the aviation sector," he says.