UWindsor researchers studying potential dangers of semi-autonomous cars

UWindsor researchers studying potential dangers of semi-autonomous cars

Over the next two years, a team of researchers at the University of Windsor will be looking into the possible dangers associated with autonomous driving.

"We hear a lot about the benefits — or the potential benefits — associated with autonomous driving. We know that they could make our cities less congested, make us better drivers, reduce crashes," explained lead researcher Francesco Biondi, an assistant professor in the department of kinesiology at the University of Windsor. 

"We're trying to figure out whether that's really the case."

Biondi, as well as engineering professors Yong Hoon Kim and Chris Lee, have been awarded $59,000 from the Social Sciences and Humanities Research Council of Canada to do just that. 

'There is a high risk'

Using a driving simulator, they will study how drivers respond to autonomous technologies, like functions that take control of the steering wheel if drivers drift outside of their lane. The semi-autonomous functions of the simulation have not yet been installed, but will be applied in the coming weeks. 

Eventually, researchers will measure the physiological and behavioural reactions of participants to see if the systems make drivers more drowsy or more inattentive, and to see how drivers react when the system requires them to step in and take control.

"We want to see how drivers react when this happens and we want to also develop a method of alerting the drivers to take control more safely," Lee said.

Biondi explained that this work is a continuation of his previous research, which has already taught him that semi-automation does come with risks. 

"It taught me that although these technologies are presented to us as ways to improve our safety and make us better drivers, in reality there is a high risk of these technologies actually making us worse off," he said.

Biondi explained that at this point in time, there's not enough scientific data to back up his hypothesis, which is what he's hoping to achieve with this project. 

Not enough evidence

Biondi hopes the data will be of use on local, national and international levels when it comes to road safety, with the ultimate goal of informing the community at large of the potential risks. 

Katerina Georgieva/CBC

"I think that right now, with the technology we have available ... we should be concerned. We're being treated as guinea pigs because there's no sufficient evidence showing that semi-automated technology is actually good."

He pointed to examples of crashes involving Tesla vehicles, and videos of drivers sleeping behind the wheel, and said, that's "very scary."

"We were promised that [this technology] would save us all and make our roads safer, but ... right now, we're seeing the opposite. We're seeing these systems being designed improperly and therefore being utilized improperly by motorists," he said. 

'Humans and automated systems don't mix well'

The ideal solution, according to Biondi, would be to either stick to fully manual cars, like the kinds of vehicles that are the norm right now, or wait until fully-autnomous vehicles have been perfected decades from now.

Anything in between, he said, where these technology systems are co-sharing control of vehicles with human drivers, won't actually make our cars any safer.

"Humans and automated systems don't mix well."

Over the next few weeks, the team will further develop the driving simulator to imitate the functions of a semi-autonomous vehicle. Soon, they'll bring in members of the public to use the simulator as participants in the study.

The researchers will publish their data recurrently over the course of the next two years, until the project wraps up in 2021.