U of T-Waterloo research data to put self-driving cars to ultimate test : Canadian winter


Do you hate driving in the winter? With the help of game-changing new data, collected by researchers at the University of Waterloo and the University of Toronto, you might never have to again.

Researchers at UW, UofT, and ScaleAI have collaborated to generate the Canadian Adverse Driving Conditions data set that will enable self-driving cars to tackle the snowy Canadian terrain.

“There are lots of great training datasets out there already, but they were collected on sunny, summer days,”  Steven Waslander, an Associate Professor at UofT’s Institute for Aerospace Studies in the Faculty of Applied Science & Engineering, said. “If you take algorithms trained on those data sets and try to use them in adverse conditions, they tend to get confused. They can misclassify objects – such as pedestrians and other vehicles – or, even miss them entirely, all because of the changes in sensor data caused by snowfall.”

The data set was collected with a Lincoln MKZ hybrid dubbed the “Autonomoose”. The vehicle – equipped with sensors, eight cameras, a lidar and a GPS tracker – collected data from more than 1,000 kilometers of driving.

Through the use of computer and human image recognition, the San Francisco-based company -Scale AI- was able to identify more than 178,000 vehicles and 83,000 pedestrian instances.

“Data is a critical bottleneck in current machine learning research,”  Alexander Wang, Founder and CEO of ScaleAI, said. “Without reliable and high-quality data ,that captures the reality of driving in winter, it simply won’t be possible to build self-driving systems that work safely in these environments.”

The next step is to implement the data set, within current software, to program self-driving cars for a Canadian winter environment.

“We’re hoping that both industry and academia go nuts with it,”  Waslander said. “We want the world to be working on driving everywhere, and bad weather is a condition that is going to happen. We don’t want Canada to be 10 or 15 years behind simply ,because conditions can be a bit tougher up here.”

This game-changing data will be critical in furthering research in the developing field of autonomous driving.

“We want to engage the research community to generate new ideas and enable innovation,”  Krzysztof Czarnecki, a Professor at UW and Researcher for the Canadian Adverse Driving Conditions data set, said. “This is how you can solve really hard problems; the problems that are just too big for anyone to solve on their own.”


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.