Custom AI models help MWAA deliver better airport experiences

Anyone who has waited in line for security at an airport can appreciate what the IT organization at Metropolitan Washington Airports Authority (MWAA) has accomplished with AI in the cloud.

MWAA’s IT team overseeing pedestrian, vehicle, and gate traffic at Washington, D.C.’s two major airports has developed a real-time intelligence platform dubbed Queue Hub that harnesses analytics, AI, IoT, and cloud to make terminal operations more efficient and profitable.

Goutam Kundu, EVP and chief information and digital officer of MWAA, says the platform, which earned MWAA a 2025 CIO 100 Award for IT leadership and innovation, was built in house for the land-side operations of both airports in the nation’s capital.

Queue Hub, begun in earnest in July 2023, was created using machine learning (ML) algorithms, open-source object detection algorithms, and custom AI models. The homegrown platform, now fully operational, runs on AWS and is enhanced continuously, Kundu says.

“It’s a living system. It’s very fluid. Things change all the time so we needed a platform that could be agile, intelligent, and nimble, and that’s why we decided to build it ourselves,” Kundu says. “It’s a very strategic project for enhancing predictability and efficiency at the airports and is a key differentiator in this market.”

The organization did not use out-of-the-box tools or programmable AWS models because there was no “cookie cutter” commercial product that could deliver the wide range of needs for the two major airports, which can exchange data using Queue Hub.

“We’re trying to get people out of queues and spend more time inside our airports, which helps boost non-aeronautical revenues,” Kundu says. “It’s the nerve center for how we take actionable insights. It doesn’t replace humans, but it does bring together our cross-functional collaboration [systems] in real-time.”

Two of the biggest pain points for travelers — enduring wait times getting though security and customs — have been alleviated directly as a result of Queue Hub’s data flow between airports and within Reagan National and Dulles International airports, which accommodate more than 50 million passengers annually.

Queue Hub, for instance, is integrated with TSA and Custom Border Patrol systems and enables demand-based resource allocation to automatically deploy resources to the right lanes at the right time to further minimize bottlenecks.

Overall, Queue Hub has reduced wait times by an estimated 15%, MWAA reports, and Dulles has one of the shortest security wait times — on average 10.5 minutes — of all major airports. Since becoming operational, Queue Hub has reduced the wait times through Customs by an average of 20%.

Addressing bottlenecks from garage to tarmac

The Queue Hub curb-to-gate data engine — which has also alleviated lines in parking garages — has made the traveling public’s experience much better and airport operations much smoother, Kundu maintains.

The platform, for instance, monitors the flow of pedestrian and vehicular traffic entering and exiting the airport and provides forecasts to ease congestion at the curb and at the gates.  

Aside from parking occupancy optimization and traffic flow management, Queue Hub aids in the management of mobile lounges at Dulles to help passengers make their connecting flights on time, which was previously a manual process.

“It gives us real-time end-to-end visibility into the curbside, garage, and terminal and helps us manage and predict congestion and passenger flows, curbside checkpoints, and parking activities,” Kundu adds. “It doesn’t replace our work but really helps us to transform how we respond to bottlenecks.”

Queue Hub also monitors and manages the flow of vehicles on the tarmac, but it is not integrated with the FAA systems or air-based systems managing airplane traffic and management, according to MWAA.

The in-house advantage

Queue Hub was built using ML algorithms and trained custom AI models on the cloud using a “hodgepodge” of IoT technologies, edge AI, computer vision, and thousands of sensors and cameras that produce “situational awareness,” Kundu says.  

The expanding collection of data from cameras at Metro stations and sensors within terminals make operations at curbside checkpoints more efficient, while also reducing TSA wait times, anticipating and resolving equipment failures, and helping passengers manage flight delays, MWAA maintains.

“When we built the AI and ML model for Queue, we use computer vision modeling and forecasts and connect it with the rest of the [IT] ecosystem,” Kundu says, adding that flight schedules, airline staffing, bagging handling systems, and biometric data all plug and play into the ecosystem.

The use cases keep growing. “We are training the models for seasonal, hourly, and event-based patterns, holiday surges, and weather disruptions,” the CIDO says. “It collects raw data from all these sources based on context, harnesses it in real-time, and we put together these models on top of it to deliver a smarter, faster, and seamless experience for our partners.”

Aside from data collection pipelines and recommendation models, Queue Hub is connected to myriad other platforms to generate positive outcomes for passengers at Washington’s two major airports, Kundu says.

 “We’ve got biometric systems, display systems, gate management systems, surface management systems, and all these ecosystems need to play together well in the sandbox,” he says.

Although Queue Hub is envisioned one day to improve runway availability and speed up aircraft turnaround times to improve aeronautical revenues, it is entirely land-based to date.

Analysts say the deployment of more advanced cloud, AI, edge, and sensor technologies at airports is on the rise.

“There is a definitive pivot towards leveraging AI and predictive analytics to tackle the historically challenging land-side equation — from chaotic curbsides to unpredictable terminal queues,” says Dave McCarthy, a research vice president covering cloud and edge services at IDC.

“Airports are rapidly recognizing that investing in these intelligent, data-driven systems is no longer a luxury, but a necessity to enhance passenger throughout, optimize resource allocation, and deliver a more seamless, less stressful journey,” he says.

Atlanta’s Hartsfield-Jackson International Airport, under the leadership of IT Director Jon Pruitt, is also putting ML and generative AI to work to digitally enhance operations based on a range of data inputs throughout the airport.