AI-rside efficiency

Airports around the world are struggling to keep pace with rising air traffic, and the pressure on air traffic controllers has never been greater. In busy towers, controllers must manage dense, fast-changing traffic on runways and taxiways, while simultaneously juggling unpredictable schedules, variable aircraft behavior, and the constant risk of congestion on the ground.

Traditionally, efforts to improve airside efficiency have focused on optimizing runway assignments, but these decisions do not happen in isolation. A choice that speeds up departures, for example, can unintentionally create cascading effects that lead to bottlenecks on taxiways or increase the likelihood of conflicts between aircraft. Few tools currently capture these interconnected effects or account for the specific operational uncertainties, such as varying taxiing speeds and schedule deviations, that controllers must manage daily.

A new study published in the Journal of Computing in Civil Engineering aims to ease that burden, while increasing airport safety, efficiency, and sustainability. The team introduces a trajectory-based simulation and multi-objective optimization framework designed to help controllers make safer, more efficient runway assignment decisions.

 “Every runway decision is a trade-off among safety, efficiency, and environmental impact,” said Pingbo Tang, associate professor of civil and environmental engineering at Carnegie Mellon University and co-author of the study. “By turning real taxi trajectories and delays into a multi-objective optimization problem, we can help controllers see those trade-offs clearly and choose options that reduce conflicts, cut fuel use, and keep aircraft moving.”

Using data from Los Angeles International Airport’s (LAX) four runways, 17 ramp areas, and more than one thousand taxiway nodes and links, researchers incorporated real-world uncertainties and evaluated their strategy during some of LAX’s busiest 15-minute periods. Across departure-heavy, balanced, and arrival-heavy scenarios, their approach consistently outperformed the first-come first-serve method commonly used in airports today, reducing taxiway conflicts by 75%, cutting fuel consumption by 15%, and increasing hourly throughput by nearly 20%.

Travelers aren’t the only ones who reap these benefits; increased efficiency also positively impacts the individuals working in the tower. Controllers spend long hours scanning complex display screens, searching for safe and efficient solutions in real time. Tang notes that while some tasks, like interpreting context or managing unexpected events, are best suited for humans, others, such as analyzing large datasets, could be completed by artificial intelligence. The goal is for humans and AI to collaborate, helping narrow the search space, make decisions more quickly, and reduce the mental workload that can lead to errors and delays

“When we design AI for airport towers, our first question is: how can this reduce mental workload instead of adding to it? If AI can filter noisy data, surface a few high-quality options, and explain why those options are safer or greener, then controllers can focus their attention where human judgment matters most,” explained Tang.

By coordinating runway and taxiway decisions and accounting for uncertainty, the study offers airports a scalable way to reduce delays, lower emissions, and improve safety without costly infrastructure expansions.

Future research from Tang and his co-authors will extend the model to include gate operations, pushback coordination, and mixed-mode taxiing, bringing these tools closer to real-world use in airport towers and supporting controllers as they navigate increasingly complex skies.