Racing tiny cars using only Artificial Intelligence and Machine Learning

Thirteen university students from across Canada are in Ottawa to put their artificial intelligence skills to the test.

It's called the Amazon Web Services DeepRacer League, where small 1/18th scale cars are being trained to complete a racetrack as fast as possible, by themselves.

“It has major components in order to do the autonomous driving," says Amanda Foo, DeepRacer Senior Technical Program Manager.

They are driven by what is called reinforcement learning.

“It’s just like training a dog," Carleton University mechanical engineering student Masoud Karimi says. "When the car performs well on the simulation we give it a good reward. And when it goes off the track, we penalize it."

The cars have a built-in HD camera to give it a view of the road and gather the information it needs to make its own decisions.

"It gathers all of its information from this camera here and determines the colour schemes, determines where’s out of bounds versus inbounds versus the middle of the track," says Foo.

"It will take a photo every 1/15th of a second, and it will analyze that photo," University of Calgary software engineering student Aleksander Berezowski says. "And it will choose to either speed up, slow down, increase going left, or increase going right."

The DeepRacer League is the world's first global autonomous racing league, open to anyone. This is the first year they have created a student league. 

"Students in high school, up to the university and college age can all compete in the virtual league,” says Nicole Foster, Amazon Web Services Director of Public Policy. "And there are prizes and opportunities for them to also earn scholarships as part of the competition."

The winning student gets a $1,000 Amazon gift card to go with their first place trophy.

But beyond the prizes, many here are looking forward to a career that helps to automate the world around us.

"Played around with some code, saw my car race around the track and then saw it improve," says British Columbia Institute of Technology student Princeton Dychinco. "So it was really cool to see the impact of what I was writing, to what could actually happen."