UT researchers seek to alleviate traffic with artificial intelligence

Researchers at the University of Texas are using artificial intelligence to help Austin's traffic problems.

AUSTIN, Texas -- We all know Austin's traffic problem is not artificial – but real.

However, artificial intelligence is working to help researchers and engineers create better traffic patterns.

The University of Texas Center for Transportation Research and the Texas Advanced Computing Center engineers starting developing the program this summer.

So far it has tracked thousands of cars and pedestrians.

They use the City of Austin’s existing traffic cameras that log hours of footage.

But having someone manually count cars is “time-consuming,” Natalie Ruiz Juri, a research associate with UT’s Center for Transportation Research said.

"We wrote a program that takes the different frames in the video,” she said. “Recognizes the objects, writes them into a database, and then we analyze the database to do whatever type of study we are trying to do.”

You can view the artificial intelligence tool online HERE.

Ruiz Juri said the numbers you see on the screen of the videos are ID numbers.

They're logged into a database where they can be used by city engineers to fix roads to make them safer and more efficient.

Ruiz Juri said this could help the city's Vision Zero program – to reduce injuries and deaths on roads.

"Can we see if there are pedestrians and vehicles in the same frame,” said Ruiz Juir. "Can we somehow use these techniques to do things only so far human analysts can look (whether) that was a near-hit or a near-miss."

UT said its artificial intelligence tool is 95 percent accurate. Researchers in the future plan to use it to identify how drivers react to pedestrian walkways and where people cross the street outside of those walkways.

This tool is still in its early stages. But the city said it is a powerful tool to fix problem locations.

UT created THIS YouTube video explaining the artificial intelligence tool.

© 2018 KVUE-TV


JOIN THE CONVERSATION

To find out more about Facebook commenting please read the
Conversation Guidelines and FAQs

Leave a Comment