Racing Robots: How the DARPA Urban Challenge Jump-Started Autonomous Tech

Posted by FF Team


The technologists piloting the next generation of mobility reflect on how a nearly 10-year-old robotics challenge served as a turning point for self-driving vehicle technology.

Autonomous driving technology has captured the attention of the automotive and tech industries, attracting billions of dollars of investment in the last few years. Autonomous vehicles are a quintessential example of disruptive technology that is changing the future course of the entire mobility sector.

However, it wasn’t long ago that R&D for autonomous driving was seen as a flight of fancy – a subject that merited quiet research in academic and internal projects, but nothing that would see any practical use outside of a lab. That was, until a critical turning point: the 2007 DARPA Urban Challenge.

In 2001, in response to the upsetting number of casualties caused by roadside bombings in foreign war zones, the U.S. Congress mandated that one-third of the military’s ground vehicles drive unmanned by 2015. The Defense Advanced Research Projects Agency (DARPA), a branch of the government focused on adapting emergent technologies, was tasked to lead this effort.DARPA reached out to several defense contractors, but when one after another failed to develop anything usable, they opted for a less conventional – and near-whimsical – approach.

They funded a contest: a robot race.

(DARPA). Winners of the 2007 DARPA Urban Challenge at the finish line.

 This sci-fi-inspired competition series culminated in the DARPA Urban Challenge – a 60-mile autonomous marathon across a multifaceted city landscape, complete with traffic signs, obstacles, and other self-driving and human-operated vehicles. DARPA enlisted academics and engineers across the world to join, and nearly 100 teams applied.

Though robot races had taken place before, this was the first time that autonomous vehicles were competing on roads that their engineers didn’t know inside and out. Cars were expected to navigate in entirely unfamiliar territory, contend with simulated traffic, and dynamically respond on the fly.

Remarkably, two key members from the top performing teams are now working with Faraday Future, pioneering our Advanced Driver Assistance Systems (ADAS) and self-driving efforts.

Hong Bae, Director of ADAS and Self-Driving.

 Hong Bae, Director of ADAS & Self-Driving for Faraday Future, oversaw the joint General Motors and Carnegie Mellon racing team. Meanwhile, Jan Becker, our Senior Director of Automated Driving, spearheaded software and vehicle testing for the Stanford University racing team.

“There were an incredible amount of unknowns,” confessed Jan. “The course, the obstacles, even the location itself was unknown!”

“It was incredibly challenging,” agreed Hong. “The development process was so rapid and intense. No matter what unexpected slowdowns arose, ultimately, race day was race day. And it wasn’t merely a race against other cars on the road – it was a race against the clock in the months leading up to it.”

On November 3, 2007, race day finally arrived. It was held at the abandoned George Air Force Base, 75 miles northeast of Los Angeles. There, a mock desert suburb was brought to life, including DARPA-hired stunt drivers to replicate local traffic. Even the hosts of TV’s Mythbusters were in attendance, narrating the event from the sidelines.

Over 100 teams were involved, some better suited for the task than others.

Jan Becker, Senior Director of Automated Driving.

 “We witnessed the first robot traffic jam!” laughed Jan.

In the end, only six vehicles succesfully completed the 60-mile DARPA Urban Challenge course.

“We were the first to cross the finish line,” Jan reminisced. “All of us on the Stanford University team could barely believe it!”

“The Carnegie Mellon team ended up starting a bit late in the day,” admitted Hong. “We were having some difficulties – our car’s start chute was close to a jumbotron TV set up to broadcast the event, which ended up causing interference with onboard components. Once we got t the bottom of the issue, we were off. Like I said: race day is race day.”

(DARPA). First place finisher, nicknamed “Boss,” earned best track time.

Though the Stanford University vehicle, nicknamed “Junior,” finished first, Carnegie Mellon’s vehicle, “Boss,” ended up earning the best track time by a small margin. Because the challenge was judged by how quickly a vehicle could complete the challenge from start-to-finish, Hong Bae and his team were awarded the winning title and project funding.

“It was a great honor, winning the DARPA challenge,” Hong confided. “It was also the first time I worked on a vehicle that could actually obey California driving rules. It could yield, park, and merge on its own with relative ease – all with dated technology from almost a decade ago!”

“The first high-resolution LIDAR sensors were released just a few months prior and computer processors were many times less powerful,” explained Jan. “There were so many limitations we were all working under. Our Stanford team ultimately came in second due to the timer, but we were still immensely proud of how Junior performed.”

“Before the DARPA Urban Challenge,” he continued, “autonomous driving technology was seen as a niche scientific interest, but little more. After this race, and after seeing how intelligent and dynamic these vehicles could be, companies swarmed to explore the possibilities that this technology could offer! It was a huge turning point – from aspirational curiosity to real-world research and development.”

(DARPA). Second place finisher, nicknamed “Junior,” maneuvering the George Air Force Base course.

 “I’ve developed several self-driving vehicles in the nearly two decades that I’ve been in the field,” Jan added. “I’ve seen the incredible strides made over the last nine years since the Urban Challenge. I’ve also seen companies make mistakes when taking on this technology. Here at Faraday Future, we’re doing things right from scratch. We have our own goals, and operate on our own terms. We don’t have to compromise this technology to conform to an older brand’s established precepts. We can pursue our own vision…we can reach for something really ambitious.”

Hong Bae and Jan Becker – leading minds from these two winning teams – are no longer rivals. Now, they are working together on the Faraday Future ADAS & Self-Driving team to harness our advanced engineering test vehicles and, with our new Autonomous Vehicle Test Permit, uncover the new potential for this technology on real-world roads.

Together, they aim to tackle autonomous vehicles’ next big turning point: getting it into the hands of people everywhere.

If Hong Bae and Jan Becker’s work in ADAS and self-driving technology interests you, we encourage you to apply for available positions on their team on our careers page.

All photographs from the DARPA Urban Challenge were provided by DARPA. To view the full image gallery from the event, visit its official online archive.

Related Article
October 09, 2016
Light Detection and Ranging is the apex of mapping technology. Discover the origins, mechanics, and its uses in the development of autonomous vehicles.
read more
June 23, 2016
Hong Bae, our Director of Advanced Drive Assistance Systems (ADAS) & Self-Driving, talks about the unique possibilities and challenges of autonomous driving technology – and how an '80s TV show inspired his approach.
read more
February 23, 2017
It has speed, luxury, and elegance. Get an inside look by the numbers at some key specs for our flagship vehicle, the FF 91.
read more