A.I. Network Gives Airport Exquisite Sense of Timing

PIT deploys local startup’s artificial intelligence tech to predict wait times, trends in security lines

By Matt Neistei

Published July 29, 2019

Read Time: 3 mins

Print/Download

You know the feeling.

You turn the corner, step off an escalator or walk through a doorway at the airport and see a huge line of people waiting to get through security. You sigh and think, “How long is this going to take?”

At Pittsburgh International Airport, you can now get an accurate answer to that question, 24 hours a day, from anywhere in the world.

The Allegheny County Airport Authority recently partnered with Pittsburgh software company Zensors to incorporate artificial intelligence into its closed-circuit camera system. Using the latest technology, computers tied into that system can count the number of people in line, evaluate their progress and calculate just how quickly (or slowly) the line is moving.

“The AI’s able to look at a frame and identify that these objects are people, and these objects are suitcases and these objects are not people,” said Zensors founder Anuraag Jain. “And then we can track the rate of flow off the change in (the number of) people.”

The software continually learns from the visual input provided by the cameras while also taking into account factors such as time of day and the number of TSA workers on duty, he said. All of that data is crunched and becomes a constantly updated assessment of how long passengers can expect to stand in line once they join it.

Those times, along with arrows indicating whether they are increasing or decreasing, are displayed on screens throughout the Landside Terminal as well as on the airport’s website, FlyPittsburgh.com.

Estimates are accurate to within two minutes, and that level of precision sets Zensors apart from similar systems at other airports, said Katherine Karolick, ACAA’s senior vice president of information technology. PIT is the first airport in the world to deploy Zensors’ “unique” combination of AI and camera footage, she said.

Jain said his vision is to take the Big Data approach that is second nature to tech giants like Amazon and Facebook and apply it to the physical world.

“If you take most traditional businesses – businesses that are brick and mortar, have a large physical footprint – there’s no cost-effective way to get the same level of granularity that you get if you’re an online business,” he said. “Our vision was how to take large spaces that are analog and make them digital.”

Zensors is part of the airport’s continuing effort to integrate emerging technologies into its daily operations, maximizing efficiency and improving the overall passenger experience. The airport authority recently launched an “Internet of Things” lab, where researchers can look for creative ways to apply those technologies in an airport setting. (The Internet of Things connects computing devices to everyday objects, such as home thermostats and security cameras, allowing them to send and receive data.)

As the Pittsburgh region continues its ascent as a leader in AI and tech innovation, its airport should reflect that – and increasingly is, airport officials said.

READ MORE: New App Allows Vision-Impaired Guests to Navigate PIT

“We want to explore different technologies,” said Karolick of the new lab. “Some of our initial priorities were looking at sensors, looking at beacons and we also wanted to look at cameras. Different devices that would pull in data.”

The Zensors system is fed by six cameras at the main security checkpoint, and it evaluates wait times for the primary line, TSA Precheck and Business Class. Three more cameras at the alternate security line are expected to be looped in soon.

And the technology doesn’t stop there, said Jain, who graduated from Carnegie Mellon University before creating Zensors and spinning it off from the university.

“I think what we’re doing with the TSA could be applied to every step of the customer journey,” he said.

For example, Karolick said, imagine booking a ticket for a flight that is two months away, immediately being told what the typical security wait times are for that day and planning your trip accordingly.

Go to Top