Science Matters →

The Race to Win at Driving the Driverless Car 

Last weekend, the legendary Indianapolis Motor Speedway was again filled with the roar of engines competing in the Indy 500. After a 2020 event with no spectators, the 2021 event had a crowd of more than 100,000 race fans--the biggest sporting event since the start of the COVID-19 pandemic.

This October, Indianapolis will host another high-stakes race on its world-famous track. Competitors will come from all over the world, but only one car will emerge victorious to claim the million-dollar prize.

And no one will be at the wheel.

The Indy Autonomous Challenge is the first competition to pit full-size driverless cars against each other at the iconic speedway. The vehicles will be identical--a cutting-edge race car retrofitted for automation--but the team behind each one will have spent months perfecting its ability to outmaneuver the others.

Among the 39 teams competing will be the University of Virginia’s Cavalier Autonomous Racing Club. Madhur Behl, an assistant professor with UVA’s School of Engineering and Applied Science, is the faculty advisor guiding more than a dozen students through the process of programming a winner.

“Most of the students are from engineering disciplines,” he said. “But it’s growing now to include non-engineering students. Students from design can make our cars look great, and business students can help us with logistics and operations. So it’s like a university-wide activity.”

UVA students
Photo: Madhur Behl

Behl said the race presents a new opportunity to test the limits of autonomous vehicles, leading to improved safety off the track in an increasingly automated world.

“Racing very naturally presents this setting where you are in close proximity with other cars, at very high speeds, and you have to make it to the end,” he said. “So if you can solve that problem, it will have a significant bearing on the safety of driving on a freeway or in a city.”

To replace a human driver with artificial intelligence (AI), the team has to write code that enables the car to “read” the road, make turns, avoid obstacles and handle the unexpected--all at top speed.

They’ve spent months practicing with a one-tenth scale model (“I like to say it’s one-tenth the scale, but 10 times the fun,” said Behl) that allows them to work on programming all of the split-second decisions of a race car. They’ve gathered for “hackathons” to troubleshoot, and will continue to fine-tune their work through simulations.

Then, this summer, they’ll travel to Indianapolis for the first of several “track days,” in which they will finally put the full-size vehicle through its paces on the 2.5-mile race track.

“The track days are our chance to tune the performance of the car for this race track,” said Behl. “There are a lot of track-specific things we have to account for in our code, so you have to make sure you understand how your car operates in these conditions.”

Behl said despite the competition for a million-dollar prize, there has been a lot of cooperation among the teams involved.

“All the universities are academic institutes and researchers,” he said, “so there is some sense that there are common problems that every team has to solve. So there is some collaboration also happening.”

UVA Students
Photo: Madhur Behl

Behl said the collective work of all the teams will contribute to better, safer driverless cars for everyday use. But he understands the skepticism of the average driver on that topic.

“I drive a stick shift myself,” he said. “I really relate to someone who says, ‘Why do we need this, I enjoy driving, it’s a stress buster.”

Still, Behl said, the call for autonomous vehicles is getting louder--and the world has been through a similar revolution in everyday transportation before.

“A hundred years ago we transitioned from horse-driven carriages to horseless,” he said. “And now we are going from driver to driverless. It’s kind of the same thing.”

Whatever happens on race day, Behl said the team will bring their car back to Virginia when it’s over. He hopes to set up a demonstration at the Virginia International Raceway so the team can show what happens when teamwork and engineering take the wheel.

“I'm excited by how this will shape the future of the role of AI and motorsport engineering,” Behl said. “I like to say that we’re just bringing AI up to speed.”