Pittsburgh’s deployment of AI won’t make security queues shorter, but it will certainly set passenger expectations and could boost airport spend

12 August, 2019

There have been lots of positives for Pittsburgh International airport since US Airways dropped the airport as a hub in the middle of the last decade. Now, 15 years on and just like the famous steel city's own business renaissance, the airport has witnessed record demand growth having successfully rebuilt itself from a hub operation to supporting what has become a booming O&D market.

There have been a few blips along the way; in the past year WOW air collapsed and Delta Air lines ended its Paris connection this summer, a route that for a long time was its sole non-stop connection to Europe. But, British Airways has returned a regular London link and Condor maintains a seasonal Frankfurt flight.

After year-on-year traffic rises of +8.2% in 2017 and +7.5% in 2018 levels have slipped to +2.4% for the first five months of 2019, still the third highest rate this decade and will move the airport close to the 10 million passenger milestone for the full year.

With Pittsburgh emerging has a tech-centre, the airport is also increasingly working to incorporate technology innovations into its campus to benefit it increasing footfall. One such addition through a partnership between its operator Allegheny County Airport Authority and local software firm Zensors, is using camera infrastructure and artificial intelligence (AI) to educate passengers to potential security check delays, helping adjust expectations.

Turning existing cameras into smart IoT sensors, the business, spun out of Carnegie Mellon University - the birthplace of AI - is able to provide actionable business insights. With the feeds piped into its general-purpose, AI-vision pipeline, its software is able to report the state of the environment with human-level accuracy.

In Pittsburgh International's case this will provide a live update on estimated wait times at the airport's Transportation Security Administration (TSA) checkpoints. This information is posted on the airport's flight information screens and website, including predictions about whether wait times will increase or decrease.

The system is working for all three lines at the airport's primary security checkpoint, and wait times at the alternate checkpoint will be added later once cameras are installed.

CHART - After year-on-year traffic rises of +8.2% in 2017 and +7.5% in 2018 levels have slipped to +2.4% for the first five months of 2019, still the third highest rate this decadeSource: CAPA - Centre for Aviation and Pittsburgh International airport reports

Pittsburgh International is the first airport in the country to utilise Zensors' technology in measuring line length. "We're constantly looking for ways to improve the passenger experience, especially with the innovative use of emerging technologies developed right here in Pittsburgh," says airport CEO Christina Cassotis. "We know security can be a frustration for travellers and having accurate wait estimates can help set expectations and aid in planning trips."

Zensors wants to "enable smart and reactive spaces through cutting-edge computer vision technologies". It believes that advances in AI should be accessible to everyone, not just those with a degree in computer science, and applied to everyday problems to make customer experiences more delightful and employees more efficient.

It uses a specially trained neural network that continually observes and learns how long passengers are waiting in line while also weighing factors such as the time of day and the number of TSA agents on duty. Estimated wait times are updated every minute, and arrows indicating whether those times are increasing or decreasing are also displayed. The company says an accurate estimate can be calculated within approximately two minutes.

The project is the latest example of the airport embracing cutting-edge technology to revolutionise the air travel process. Partnerships with dynamic firms like Zensors will continue to improve operations that directly affect passengers, employees, tenants and airlines every day.

Anuraag Jain, head of product at Zensors, says the project is the just beginning of leveraging AI in the airport environment. "Airports are a perfect use case for this type of technology. We're applying deep learning in a way that can really become a game-changer for passengers and airport operations," he says.

Alongside setting customer expectation levels, the new system has obvious customer service benefits and could also boost traveller spend. A view on the expected queue time at security means more free time in the terminal. Additionally, the process will be less of stress and we all know a happy traveller is more inclined to spend.