Transurban is using artificial intelligence to help it bill toll road users whose e-tags aren’t detected or recognised as the vehicle passes under a collection point.
Artak Amirbekyan, Transurban
The toll road operator has built an “auto-correction” AI model to identify vehicles missed by its e-tag system.
Operational on Transurban’s toll roads across Australia, the AI model analyses imagery from automatic license plate recognition (ALPR) cameras to identify a car’s make, number plate and location, which is then combined with other customer data to determine where to send an invoice.
This, according to Transurban’s head of data, AI and ML, Artak Amirbekyan, has reduced the number of images sent for human review by up to 40 percent.
Speaking during Gartner Symposium in the Gold Coast, Amirbekyan said that “most customers use e-tags” on its roads, which for the most part “work very well.”
“But sometimes the tag doesn’t work; someone doesn’t have it or [there have been] some changes. So, we need to read the licence plate to be able to identify the car and the customer.”
Transurban processes roughly 2.5 million customer trips each day across Australia and most will be billed through the e-tag system.
Vehicles that aren’t picked up by an e-tag are photographed by Transurban’s ALPR cameras. Images that are hard to read are passed to Transurban’s team of agents to piece together the vehicle’ identity and bill the driver.
AI-enhanced ALPR uses a machine learning platform built on Amazon SageMaker, which is able to identify the vehicle with 99 percent accuracy, according to Amirbekyan.
The high degree of accuracy was important, he added, because “otherwise you will get complaints” about incorrect bills.
A long-term user of data and analytics, Transurban has been embedding AI into its organisation using Amazon SageMaker and, more recently, DataIQ.
Its data, AI and ML team is now running a number of uses cases alongside AI-enhanced ALPR, including road safety and incident detection and tunnel ventilation.
Speaking at Gartner Symposium alongside Amirbekyan, Transurban CTO Tanya Trott said almost 40 percent of the company’s staff work in tech-related roles.
“While we are one of the safest road operators in the world, there’s still a lot more we think we can do to improve safety, and data is playing an integral role in our ability to do that,” she said.
Eleanor Dickinson attended the Gartner Symposium as a guest of Gartner.