AI being tested to detect distracted drivers in Canada

For the first time, artificial intelligence used to detect distracted drivers is being tested in Canada.

The system, which uses several cameras to capture high-resolution images through windshields before using AI to detect if a device is being used by the driver, went live in Edmonton Tuesday.

University of Alberta engineering professor Karim El-Basyouny is overseeing the three-week pilot project, which is specifically looking for those using devices while driving. Although this same technology is used for ticketing in other countries, the Urban Traffic Safety Research Chair stresses this project has no enforcement attached.

“Data around distracted driving in Canada is a little patchy,” says El-Basyounym “We don’t understand how bad of a problem it is.”

Current data around distracted driving comes from ticketing, which largely happens when officers observe someone using a device at a red light. This system can capture if a device is being used while a vehicle is in motion.

“We know that if you are slightly distracted that is going to increase your perception and reaction time, which then leads to longer distance to stop, inability to react to the hazards on the road.” Says El-Basyouny.

While this is the first time the technology developed by Acusensus is being tested in Canada, it’s already being used in other countries for enforcement.

“In Australia, there’s a reduction in fatalities in states that have our technology,” says Tony Parrino, the general manager for Acusensus in North America. “They attribute that to the awareness from the media about the dangers of distracted driving, and through the fact, people are getting repercussions for their actions.”

Parrino says their system has also been used for real-time law enforcement in parts of the United States, by “sending information down to a police officer, who will get information that the vehicle is approaching and pull people over and write them a ticket.”

Although the AI used for this research project is only looking for drivers using devices like cell phones, Acusensus has the ability to detect if seat belts aren’t being worn and even signs of impaired drivers.

This research is funded through a joint initiative by the University of Alberta and the Edmonton Police Services, and the findings are expected to help the city in future decisions on whether to use similar technology in Edmonton.

“It’s important that we continue to evolve to meet the changing needs of Edmontonians by supporting safe streets for all,” says Jessica Lamarre, director of Safe Mobility for the City of Edmonton in a news release.

For El-Basyouny, he knows police can’t be everywhere, saying “enforcement for distracted driving strains police resources, is difficult to detect, and is subject to possible bias from officers who must interpret behaviour and body position to determine an offence. Automated technologies can reduce these issues and provide a more equitable enforcement method.”

In order for this technology to be used for enforcement, it would require a change in provincial laws.