University of Waterloo team trains autonomous car to drive in winter weather

As motorists across much of southwestern Ontario endured poor road conditions Tuesday, the "autonomoose" - a driverless Lincoln MKZ - is also getting a lesson in winter weather driving.
The Renesas autonomous vehicle and Autonomoose, the uWaterloo autonomous vehicle, are seen here operating under winter driving conditions in Stratford in December 2016. (Elke Bidner/investStratford)

As motorists across much of southwestern Ontario endured poor road conditions Tuesday, the "autonomoose" is also getting a lesson in winter weather driving. 

The University of Waterloo's Centre for Automotive Research is one of three groups that got the green light in November to test autonomous vehicles on Ontario roads

Its Lincoln MKZ, also known as the "autonomoose," spent two months learning to drive in icy, snowy conditions in Stratford, Ross McKenzie the general manager of the centre, told CBC News. 

"We created a test facility to drive autonomously with the challenging conditions with snow and ice and didn't have the benefit of having the roads salted so we could test in true strenuous conditions."

And like human drivers, the computer-driven car also struggled with the rough weather. 
University of Waterloo graduate student researchers Carlos Wang (left) and Ian Colwell track autonomous driving operation of the Renesas vehicle under winter driving conditions in Stratford in December 2016. (Elke Bidner/investStratford)

"We wanted it to be challenging and we wanted to learn a few things."

One of the things they learned, said McKenzie, was the lidar system (a laser-radar sensor) misread ice when detected on the road. 

"We discovered that when the laser beam hits ice it refracts differently, as opposed to just reflecting straight back off of the pavement," he said.

"It actually gave a reading that the ground was lower than it actually was," similar to a pothole, McKenzie said. 

Not a deal breaker

The glitch is far from a deal breaker. In fact, the WatCAR team will use those errors to teach the cars how to better respond to those conditions.

"It's the exact opposite of a deal breaker – it's what we want. We want to experience those incidents so we can program the computer on the vehicle to expect them and react appropriately."

The circumstances were then repeated "hundreds if not thousands" of times "and that allows the vehicle computers to make an informed decision when it encounters that situation on the road," McKenzie said. 

Once the vehicle learns how to handle these conditions, it will be safer on the road than any human-controled vehicle, he said. 

"Some people don't check their blind spot. Other people change lanes without signaling and everybody's late at times and either goes over the speed limit or, you know, pushes an intersection to get through when it's amber and mistakenly misses intersections and runs red lights," he said.

"And an autonomous vehicle will never do any of that."


To encourage thoughtful and respectful conversations, first and last names will appear with each submission to CBC/Radio-Canada's online communities (except in children and youth-oriented communities). Pseudonyms will no longer be permitted.

By submitting a comment, you accept that CBC has the right to reproduce and publish that comment in whole or in part, in any manner CBC chooses. Please note that CBC does not endorse the opinions expressed in comments. Comments on this story are moderated according to our Submission Guidelines. Comments are welcome while open. We reserve the right to close comments at any time.

Become a CBC Member

Join the conversation  Create account

Already have an account?