Snow still a challenge for self-driving cars

VANCOUVER (NEWS 1130) – Would you like to hop in a self-driving car to avoid having to make your way around in the snow?

One expert tells us the technology continues to improve, but there’s still a ways to go when it comes to winter weather.

Assistant professor Steven Waslander with the University of Waterloo’s mechanical and mechatronics engineering department has been helping develop and test a new autonomous car for Renesas Electronics America along with his 25-person team, and things were going pretty well until the snow hit.

“What happens essentially is that the sensor data you get back is just lower quality,” says Waslander.

“There’s a lot more noise, a lot more reflection, and sometimes just no signal at all returned. That means your ability to detect the environment and the other vehicles operating near you or pedestrians is diminished when you’re operating in snowy conditions.”

Waslander is optimistic you’ll one day be able to use a self-driving car in the snow but concedes this could mean Canada’s colder regions may have a slower rollout of the technology.

“There’s definitely more work that’s going to have to be done in terms of refining the algorithms so they’ll also work in snowy conditions, but there are just as many hurdles remaining in getting to human-level performance in regular environments as well,” says Waslander, who suggests when self-driving cars do get the green light here, there could be certain conditions or areas where drivers will not be allowed to use autonomous features, at least initially.

Given the experimental car’s struggles this winter, Waslander’s team plans to gather more information in the coming weeks by doing more driving.

Then, based on what his researchers observe, they’ll try to figure out ways to improve the technology for more testing next winter.

Top Stories

Top Stories

Most Watched Today