Just west of Willows, CA in rolling hills bordering the northwest Sacramento Valley is the Thunderhill Raceway Park. The motorsports complex offers plenty, including a pair of winding road courses which can be combined into the longest road track in the U.S. Toss in a skid pad, catering services, and club houses and the park has everything expected by racing enthusiasts. Except for one weekend a year, when a once-critical component is conspicuously absent from the main event—drivers.
During a weekend in May 2016, Silicon Valley entrepreneur Joshua Schachter held the inaugural track day for Self Racing Cars (SRC), which he envisions as becoming an autonomous racing series. Generating significant attention, numerous media outlets covered the event, including Car & Driver, Forbes, TechRadar and many more (story links available on SRC about page). In an interview for the Velodyne 360 Blog, Schachter explained, “The number one question from people was, ‘When’s the next one?’”
The second Self Racing Cars track day was recently held the weekend of April 1st, 2017 at the 2-mile Thunderhill West course. Over 100 people came from about 20 large and small companies involved in autonomous driving technology, such as AutonomousStuff, comma.ai, and PolySync. Schachter explained that while the goal is to slowly make the event more competitive, currently the approach is more collaborative. “Like a conference with an engineering competition.”
Toward that end, humans first drive test laps to collect data with various sensors and systems, including GPS, LiDAR, RADAR, and optical imaging. This data is shared between teams on the SRC website to help participants develop their own autonomous systems. “The [current] race is for the engineers,” said Schachter. “It’s not like NASCAR. This doesn’t have that element of drama. The audience is the people competing. The spectators are those interested in the field.”
Still, this year’s track day saw the addition of one key racing element. “We’re figuring out how to add to that competition very gradually,” said Schachter. “We don’t want to make it too hard too fast.” For the 2015 track day, “winning” meant simply that an autonomous car completed an entire lap. A feat which was self-reported by two finishers. But for this year’s event, each autonomous car was timed during roughly 4-6 sessions. These sessions involved flying laps, which means the cars cross the starting line in motion. Of the nine autonomous contestants, four completed the course: Point One was the winner at 3:37.97, followed by AutonomousStuff at 3:44.83, comma.ai at 3:47.58, and Vector AI at 3:51.275.
Currently, each autonomous car races alone on the track during their time trial. But someday, Schachter sees the series evolving to resemble a typical car race. “We’d like to have full wheel-to-wheel racing,” he said. “But right now, the cars can’t all see each other.” What’s needed will be new innovations in autonomous software. And Schachter believes that strategically designing future courses and rules can encourage such innovations.
One option to spur development, Schachter shared, is placing obstacles on the track for the autonomous race cars to avoid, such as a fake pedestrian. Another option is including tire walls, like those featured in Formula 1 racing, which protrude into the course and necessitate nimble steering adjustments. Eventually, friction dynamics and limit-handling might come into play, which would involve developing an autonomous car with an anti-traction control. Such a feature would allow the car to slide around the curves like how human drivers will drift through a turn. “As you extend it out,” says Schachter. “[We’ll] get people building vehicles that are ever more competitive. For the racing, it’s about pushing the limits of what’s possible.”
But for now, the steps are incremental and plenty of room remains for improvement. To illustrate, Schachter references the recent track day when the best 2-mile lap by a trained (human) driver in a SR3 race car was 1:13. The fastest times for the autonomous cars were three times longer. Ultimately, Schachter—who created Self Driving Cars due to an equal love for robotics and racing—sees the track days as the path to an eventual racing series. But other participants see it as a means to spur the ongoing development of driverless technology in general.
During the 2016 event, Translogic host Jonathan Buckley interviewed several participating CEOs to understand their motives for competing. Said Robert Hambrick, founder & CEO of Autonomous Stuff, “For us, it’s about enabling automation, so it’s not about racing…but building modular-based applications that we can give to the world of researchers.” Echoing such sentiments, George Hotz, founder & CEO of comma.ai explained, “If you can understand how to race, you can also understand what to do in a lot of evasive safety-critical situations.”
John Eggert, Director of Automotive Sales & Marketing for Velodyne LiDAR, explained his company attends Self Racing Cars to “interact with a much wider part of the ecosystem, including smaller self-driving car startups, ground robotics companies, sensor and mapping companies, [and] autonomous shuttle companies.” As an example, Eggert points to participant Udacity as the premier online education training ground for autonomous vehicle engineers. Velodyne sees friendly events like Self Racing Cars as an opportunity to support the hands-on education of future leaders in the field. By bringing sensors to the event, and providing data to racers, Eggert stated that Velodyne wants to share “the benefits of high definition 3D LiDAR [with] tomorrow’s autonomous vehicle engineers.”
Meanwhile, the creator of Self Racing Cars is excited about what his racing series might engender in the broader field of driverless vehicles—even if that’s not his personal motivation. “For decades, racing has influenced [the design of] cars on the street,” said Schachter. “The seat belt. The rear-view mirror. Hopefully, the same diffusion of ideas happens from self racing cars. But personally, I just like sliding around the track.”
So, when’s the next one?
“That’s a good question,” said Schachter, who explained that, for now, he’s financing the event out of pocket. “I’m still looking for a sponsorship.”
Given the buzz surrounding Self Racing Cars, it seems quite possible a sponsored race series may be looming on the horizon.
Velodyne Lidar (Nasdaq: VLDR, VLDRW) ushered in a new era of autonomous technology with the invention of real-time surround view lidar sensors. Velodyne, a global leader in lidar, is known for its broad portfolio of breakthrough lidar technologies. Velodyne’s revolutionary sensor and software solutions provide flexibility, quality and performance to meet the needs of a wide range of industries, including robotics, industrial, intelligent infrastructure, autonomous vehicles and advanced driver assistance systems (ADAS). Through continuous innovation, Velodyne strives to transform lives and communities by advancing safer mobility for all.