Hit enter to search or ESC to close

What’s the Deal With Autonomous Racing? New Races Redefine Motorsports for the Self-Driving Era

Courtesy of Indy Autonomous Challenge

Since the 1960s, when legendary Ford dealer Bob Tasca Sr. coined the phrase, “Win on Sunday. Sell on Monday,” major global automakers have poured billions into motorsports. Their chief reason: Auto racing shows off the latest, greatest car technology to consumers, who can close their eyes and envision being the next Senna, Schumacher, or Danica even while they’re in a bumper-to-bumper slog on the Interstate.

But that once-airtight logic is being tested, as Millennials and Gen Z consumers prioritize private car “access” over ownership. If young people, especially those living in cities, are content to get where they’re going by hailing a ride or hopping on an electric scooter, then what’s the point of elaborate experiential marketing to win sales of private cars? Why create fandoms for racecar drivers when driving itself is growing less attractive?

Those are the questions raised by three fledgling racing competitions that feature vehicles driven not by humans, but by robots. These autonomous racing series are testing an entirely new hypothesis in motorsports: that by demonstrating the technical capabilities of fine-tuned sensors and sophisticated algorithms on a racetrack, consumers will eventually warm to the idea of hopping in self-driving vehicles themselves.

“The motorsports world is trying to maintain relevance to what is happening in the passenger car world,” says Dr. Robert Prucka, a professor of motorsports engineering at Clemson University. Prucka and a group of students engineered the first prototype vehicles for the Indy Autonomous Challenge, the newest of the newfangled autono-races.

Prucka cops to being “a huge racing fan,” but he believes that such go-fast technologies as “powertrains that go 200 miles an hour” are arguably less relevant to consumers today. When nearly every new car driven off the lot includes automated braking and lane-keeping assistance, consumers are growing more comfortable with the idea that their cars aren’t dialed-down versions of hot rods, but really high-performance computers on wheels. “Autonomous racing brings all of that right back in,” Prucka says. 

Giving Autonomy the Indy Treatment

The latest of these efforts, the $1.3 million-purse Indy Autonomous Challenge, was announced in November 2019 by the famed Indianapolis Motor Speedway in partnership with Energy Systems Network (ESN), an Indiana-based non-profit. The competition is modeled on the DARPA Grand Challenge, the Defense Department’s 2004-2007 prize contest that spawned early development of self-driving cars. 

Each high-tech vehicle competing in the Indy Autonomous Challenge costs more than $1 million to produce, according to Matt Peak, director of mobility at ESN. With support from corporate and foundation grants, teams are mainly comprised of participants from top universities, including MIT, University of Michigan, and Cal Poly. Each university team buys its vehicle for approximately $300K—with the event organizer providing a favorable lease or loan if necessary.

The event’s stated mission is to spur innovation in energy technology and to foster the next generation of autonomous vehicle engineers. And while the winning team can walk away with $1 million (and second and third places get $250K and $50K, respectively), NASCAR this is not.

On October 23, 2021, the Challenge will hold its inaugural race featuring as many as 10 university teams fielding cars on the track at the same time. Each team will use an identical vehicle, the IAC AV-21built on a Dallara chassis that’s similar to what’s driven in the IndyCar series. Sponsors like AdLink, Bridgestone, Luminar, Microsoft and others will provide state-of-the-art processors and sensors, as well as traditional racing gear, like racing tires and motor oil.

The race’s main challenge isn’t to show how fast a self-driven race car can go, but rather to demonstrate how autonomous vehicles deal with hazards and respond to the unexpected.  “If the technology used on an autonomous race car at 200 miles per hour can avoid other cars, navigate around them, and not lose control,” says Peak, “it dang well sure is going to help the average driver avoid a piece of lumber that falls off of the truck in front of them on the highway.” 

The Track as Test Platform

Courtesy of Roborace

If the Indy Autonomous Challenge is more interested in incubating technologies and boosting interest in the self-driving industry, then the founders of the London-based autonomous racing series Roborace have considerably more commercial aspirations. They hope that one day autonomous racing will become a highly profitable spectator sport–think a cross between Formula 1 and Battle Bots. 

The inspiration for the series, says Roborace co-creator Mikhail Sokolov, was born in 2015, when the company was working with Formula E, the pioneering electric-vehicle racing circuit, on a series of demonstrations of autonomous vehicles. “Then we got the idea for another kind of motorsport series that could showcase super extreme autonomous robots,” Sokolov says.

The Roborace organization is a sister company to Arrival, a startup that makes commercial electric vehicles in micro-factories. The list of participants currently consists of five university-based teams, including MIT Driverless and the University of Pisa, with Arrival Racing as the single for-profit enterprise. Acronis SIT, another team, is part of the Switzerland-based Schaffhausen Institute of Technology, a non-profit research institute. (Brand-name autonomous vehicle companies have thus far taken a wait-and-see approach.)

Roborace eventually hopes to stage an entire field of vehicles racing at the same time at the world’s most prestigious racing venues. But that’s not quite happening yet. Starting with proof-of-concept events in 2016, most Roboraces have fielded only one or two cars on the track at a time. Many have failed to complete the race, due to a combination of mechanical failures, battery issues, faulty control signals, and systems miscalculating the location of obstacles. These difficulties reveal that the racetrack is as grueling and unforgiving of an environment for autonomous vehicles as it is for human-driven ones—and maybe more so. 

Roborace achieved a major milestone on April 9. On Round 9 of the current “Beta Season,” all six teams completed the circuit, one at a time, at the Las Vegas Motor Speedway. For the first time, there were no DNFs—Did Not Finish. Head-to-head competition with at least two autonomous vehicles racing on the track at the same time is slated for spring 2022 during Roborace’s official Season 1.

After teams transfer their software to Roborace vehicles—loosely modeled after LeMans cars—they cannot make changes. “They just push a button, and the car runs,” says Sokolov. “They may get logs or telemetry afterward, but they are not interacting while the race is running.”

As if that challenge isn’t tough enough, Roborace this year introduced virtual objects to the track. Computer-generated walls might suddenly appear to create unexpected obstacles. So-called ghost cars could join the race at any moment. This AR racing environment—a bit like Pokémon GO, but at breakneck speeds—is dubbed the “Metaverse,” an extension of Roborace’s patented technology for vehicle simulation and data fusion. To score bonus points (“loot” in official terms), teams pick up digital objects, jewel-like geometric tokens, that can pop up anywhere on the track—a task that requires the cars to use dynamic and responsive path planning.

Sokolov emphasizes the importance of intelligent software-based policies for braking, acceleration, and dynamic path planning. “Winning is mainly about the dynamics, how you plan for that in your code,” he says. “In our competitions, cars running at speeds at nearly 100 miles per hour are getting similar results to teams with more conservative speeds.” 

Professor Prucka of the Indy Autonomous Challenge emphasizes a similar defensive approach, one that aligns with the safety principles underpinning automated vehicle systems in today’s cars. But autonomous racers have one key difference: they’re trying to go faster than everyone else on the road. “You have to get to the finish line,” says Prucka. “So initially, the strategy is incredibly conservative.” He advises eager university teams to “give way for the first few laps and let things hopefully string out a little bit and only then start to pass.” 

The final rounds of the 50-mile, 20-lap race at Indianapolis will be a different story. Giving way? Not so much. ”When there’s a million dollars on the line and you hit the last five laps, you’re only trying to pass,” he says. 

Expensive Common Platforms vs. Grassroots Innovation

Courtesy of Self Racing Cars

Both the Indy Autonomous Challenge and Roborace go to great lengths to ensure every team is competing with the same vehicle and technology platform. Each team’s algorithm is the single differentiating factor.

However, for Joshua Schachter, the impresario behind Self Racing Cars, the high expense of exotic racing platforms and sensors can be a roadblock to innovation. Launched in 2016, Self Racing Cars’ annual races at the Thunderhill Raceway Park in Willows, Calif., bring a mixed crowd of weekend hackers competing in self-owned second-hand hybrids and electric vehicles. Teams that log the fastest laps get slaps on the back, but Self Racing Cars offers no prizes and only a handful of spectators. The event is less like a traditional auto race and more like a robotics hackathon set at a professional racetrack. “I want people to say, ‘This is a stupid idea, but it might just work,’” Schachter says. “I think there’s an enormous amount of power in riffing.”

The cars at Self Racing Cars commonly have an Arduino board—an open-source microcontroller platform—dangling from the dashboard. Xbox controllers are mounted on steering wheels, and cameras bought off Amazon provide vision. Clearly, Schachter wants self-driving experimentation to be as democratic as possible. “If you can bring down the cost, you’ll get more participation,” he says.

A Silicon Valley entrepreneur and investor, Schachter has a degree in electrical and computer engineering from Carnegie Mellon University. He was the founder of  Del.icio.us and is credited as the inventor of social bookmarks and tags.  The garage in his Bay Area house is a Wonka-esque workshop for low-budget robotics experiments, like the old-school go-kart that he recently tried to control with computer actuators. “Honestly, just starting the thing in my garage was so earsplitting that I couldn’t really test it,” he says. 

Schachter believes that revolutionary ideas can come from anywhere, especially when passionate technologists interact with each other’s work. And the lower the stakes, the better. He’s exploring a common platform of kart-like vehicles—perhaps with quickly replaceable parts in case of a collision. That would eliminate fears of crashing a million-dollar car that might inhibit taking chances. “Racing is very much about locking down the platform to ensure competition. But my real goal is not to have competition so much as innovation,” he says. 

Critics of autonomous racing might argue that extreme high-speed scenarios bear little resemblance to the trips we take to the grocery store or visit grandma for the holidays. But all three of the autono-racing series’ organizers believe that that argument overlooks a fundamental aspect of racing: the need for safety.

To win a race, after all, you need to finish the race. That’s a longstanding maxim for all motorsports, but in the new world of autonomous racing, a flaw in an algorithm could quickly lock up a vehicle’s controls and send it off the course. Algorithms that simply mash the accelerator are worthless if the vehicle’s perception system can’t avoid collisions with other robotic racecars or recognize a tire bouncing down the track. 

Eventually, successful high-speed autonomous races could go a long way to demonstrating the technology’s readiness for widespread public use. Today’s drivers have come to trust relatively basic safety systems that beep warnings if they approach another car too briskly or drift out of the lane. But, according to recent surveys, they remain wary about a self-driving vehicle’s ability to handle the countless random hazards that arise while driving.

Could those concerns begin to vanish with the sight of a crowded field of human-less race cars whizzing by one another at Indy speeds—without crashing? In the end, that’s what could make autonomous racing a winner. 

Choose your lane

How Autonomous Vehicles Distinguish Between Bicycles and People Who Ride Them

How Autonomous Vehicles Distinguish Between Bikes and People

When it comes to how autonomous vehicles see the world, humans come first, literally. Autonomous vehicles (AVs), like the kind operated by Pittsburgh-based Argo AI, use Machine Learning to detect and classify the objects in their surroundings, identifying people...
Why The League of American Bicyclists is optimistic about autonomous vehicles

Why a Leading Cycling Advocacy Group Is Optimistic About Autonomous Vehicles

As autonomous vehicle use grows, AV companies and the League of American Bicyclists are collaborating on how to ensure cyclists and motorists can share the roads safely, even if the “motorist” is artificial intelligence software. As part of the...

Self-Driving Is Arriving Right On Time. Just Like Ice Cream Did

Seven years ago, I was a self-driving skeptic. Not of the technology. Of all the “experts” promising autonomous vehicles would be everywhere by 2020. You didn’t need to be Nostradamus to know that was ridiculous. All you needed was...
Illustration of a futuristic parking deck turned into a mixed-use space, with AVs driving by

How Autonomous Vehicles Could Help Transform Parking Lots

Researchers say it’s likely that autonomous vehicles (AVs) can help reduce the need for parking lots, opening more room for grass and trees and other elements of nature. It may not seem like it when you’re circling the block...
An illustration of an Argo autonomous vehicle in teal against a backdrop of buildings, a bicyclist, and research papers

7 Big Breakthroughs From Argo Research at CVPR 2022

The 2022 Conference on Computer Vision and Pattern Recognition (CVPR 2022) is nearly here. Thousands of computer scientists, software engineers, and researchers from around the globe will gather in New Orleans to review and discuss their latest work in...

Researchers Predict the Future With Lidar Data

Researchers at Carnegie Mellon University’s Argo AI Center for Autonomous Vehicle Research, a private-public partnership funded by Argo for advancing the autonomous-vehicle (AV) field, say they have come up with a way to use lidar data to visualize not...

Must Reads