Artificial city

Distracted drivers are identified by artificial intelligence in Edmonton

Artificial intelligence is currently being used in Edmonton to detect distracted driving as part of a research project.

On September 13, the University of Alberta launched this three-week research project to understand the prevalence of distracted drivers, particularly in Edmonton. Karim El-Basyouny, a professor in the faculty of engineering and holder of a research chair in urban road safety at the University of Alberta, leads the research team. U of A research is in collaboration with Acusensusthe City of Edmonton and the Edmonton Police Service.

As of September 13, the technology has been parked at its first location at the intersection of 79th Street and Argyll Road. According to El-Basyouny, he will be stationed there for about a week before moving to the next location, which is currently unknown. There will be a total of three different locations, one for each week during this project.

El-Basyouny’s research is supported by a seed grant, which makes the use of Acusensus technology possible. Although the Edmonton Police Service is collaborating on this project, data collection will be used for research purposes only, not traffic enforcement.

Edmonton is the first city in Canada to test Acusensus technology, according to Tony Parrino, general manager of Acusensus in North America.

“The data on distracted driving in Canada is a bit patchy, [and] we don’t really understand the magnitude of the problem… what we’re trying to do is see if there’s a better way to understand the magnitude of the problem [distracted driving] east,” El-Basyouny explained.

Lily Polenchuk Karim El-Basyouny (left) and Tony Parrino (right) standing in front of technology used to identify distracted drivers on 79 Street and Argyll Road.

Acusensus AI technology used to collect data

The technology used to determine the prevalence of distracted drivers is primarily AI. According to Parrino, the AI ​​went through “a number of training scenarios with millions of data points.”

The system is radar-based with many different sensors and four different cameras. Each camera captures something different; one captures a steep shot of the windshield, one camera is shallow for a phone-to-ear event, and the other two cameras are used for color context and license plate capture. The collected information is then transmitted to the AI.

According to Parrino, although the AI ​​was trained for maximum accuracy, there is a possibility of false positives.

“It’s very accurate, but there are false positives… 100% of captured images are reviewed by trained people [who determine if] the criteria are met for the U of A to determine that a distracted driving event has occurred, and only those are counted,” Parrino said.

Future use of this technology in Edmonton is unknown

Although Acusensus technology is being used in Australia for traffic control, according to Parrino, it’s unclear if the technology will be used for traffic control in Edmonton. Currently, this research is only used to see the prevalence of distracted drivers in Edmonton.

“I think [traffic enforcement] is an option available to us at [some] point in the future, [however] that is not primarily the purpose of this study,” El-Basyouny said.

In a statement sent Sept. 13, Jessica Lamarre, director of Safe Mobility for the City of Edmonton, commented on the University of Alberta research project.

“This project provides an opportunity to better understand the prevalence and safety effects of distracted driving on our streets through the creative use of new technologies alongside our talented research partners at the University of Alberta.