Winning once is difficult; continuing to improve after many years at the top is even harder. But U of T’s self-driving car team — aUToronto — has done just that, placing first at the final SAE AutoDrive Challenge competition, their eighth win in nine years.
“We expect both the leads and the members to deliver a lot more than what would normally be asked of a student design team,” says Chad Paik, a PhD student at UTIAS and the team’s captain.
“The bar is set really high, and because several people from last year decided to rejoin this year, we were able to keep momentum, which was really helpful.”
Each June for the past several years, the team has competed against nine universities from across North America at the Mcity Test Facility in Ann Arbor, Mich. Sponsored by General Motors (GM), 2026 marked the last year of the AutoDrive competition in its current form and the final year of participation for aUToronto.
The rules and parameters around the competition change each year, ensuring teams are continuously tested. This year, one of aUToronto’s goals was to improve their score on the localization challenge, one which has them navigate their vehicle to waypoints with a weakened GPS.
“Originally, we tried out lane localization, something that we thought would improve our performance. But after developing the algorithms and testing, we discovered it didn’t provide the centimetre-scale accuracy we needed,” says Cecilia Li (Year 3 EngSci).
“We pivoted to different state-of-the-art LIDAR approaches around six weeks before the competition, which worked, but the approach was completely different. We weren’t afraid of making changes and embraced the discomfort that came with them.”
Their risk paid off. The team scored full points, achieving the only perfect score in the localization challenge and improving their score by 130 points over last year.
One of the big surprises for the team this year was how smooth the competition went.
“I’ve been to the competition the most out of everyone, and when we go there’s usually always a major surprise or a big issue that comes up where we’re very stressed,” says Paik.
“That’s not to say that we didn’t have any issues this year, but we had fewer and more predictable problems, which was new. Everything went as smoothly as one could hope for.”
With the end of the SAE AutoDrive Challenge, the team is already turning its attention to a new goal. They plan to autonomously drive downtown to the St. George campus from UTIAS in North York. Unlike the closed-course AutoDrive competition, the project will require their vehicle to navigate real city streets and traffic.
“Usually, we take a month break after the competition before we start back up again, but what we’re trying to do next is so difficult that we basically gave everyone a weekend off before we were back at it,” says Connor Wilson (Year 4 EngSci).
They have several hurdles to work out in preparation, one of them being the weather.
While self-driving cars are already being used as robotaxis in 10 U.S. cities such as Austin, Tex. and Phoenix, Ariz., driving in colder climates remains a hurdle for autonomous vehicles.
The team has some experience with this as the competition has included inclement weather elements they’ve had to anticipate, such as driving through fog and against different lighting and glare.
Over the next year, they’re hoping to further their work and collect a data set focused on winter driving behaviour prediction to teach autonomous cars how people and traffic behave in snowy and icy conditions.
“Winter conditions are more difficult and dangerous, in part because you lose traction and visibility, and in the case of autonomous vehicles, sensor information quality,” says Professor Steven Waslander (UTIAS), academic advisor to the aUToronto team and Director of the Robotics Institute.
“The team has done some fantastic research over the years and it will be great to see what they produce given this new challenge.”
The relationships formed as part of aUToronto have translated into professional opportunities for many of the members.
Li is completing her PEY Co-Op with GM. She says many members of the team have gone on to work for the car company in part due to their experience with aUToronto.
Wilson, meanwhile, was recently doing work with the NBA in Las Vegas as part of his PEY Co-op placement with Peripheral Labs, a computer vision startup founded by a former aUToronto member.
“I think the team here at Peripheral is almost half aUToronto alumni,” says Wilson.
“Everyone keeps joking that this is like our competition week, partially because we’re working 24/7. I definitely think aUtoronto prepared me to work super hard.”
“I can proudly say we achieved all of our goals for this year, from redeeming ourselves in the localization challenge to winning again despite more technical tests,” says Paik.
“The technology around autonomous cars has changed so much from when the team started and we’re ready to put it to use and try something new.”
