Stanford rises to academia’s Big 3 of autonomy
Palo Alto, Calif. — They engineer what seems like magic under the roof of an unassuming seven-bay garage on the west side of Stanford University’s main campus.
A driverless DeLorean capable of drifting. Solar-powered and self-driving race cars. A test vehicle capable of monitoring driver brainwaves as well as simulating the effects of moving on ice.
They call it the Automotive Innovation Facility, and it has become a focal point for hands-on research linked to autonomous, electric and connected vehicles. It’s an arm of the Center for Automotive Research at Stanford, where students and faculty have established the private university as one of the three main education centers sparking innovation in the field today.
Researchers here partner with a variety of companies in the tech and auto industries — all working to make driverless vehicles safe and bring them to the masses. It’s a similar role currently being played by the University of Michigan and Carnegie Mellon University in Pittsburgh, constituting what some experts see as academia’s Big Three in autonomy.
“We have students here who have ideas in their heads that we think can change mobility,” J. Chris Gerdes, director of Stanford’s Dynamic Design Lab, said in a wide-ranging interview with The Detroit News. “And we think those ideas will sort of stay in their heads unless we give them the opportunity to bring them out on the road.”
Pittsburgh’s Carnegie Mellon, long known for its top-flight robotics program, traces its involvement in self-driving technology back more than three decades. Its project partners include General Motors Co. and Uber Technologies Inc., with the latter having hired away several of the university’s researchers.
CMU’s latest autonomous vehicle is a 2011 Cadillac SRX that “takes ramps, merges onto highways and cruises at 70 mph.” Current research centers on improving how data collected by car sensors is analyzed.
“Algorithms process this information, which is used to control the vehicle, so it is imperative that environmental data is interpreted correctly,” Abby Simmons, associate director of media relations at Carnegie Mellon, wrote in an email. “This work will help autonomous vehicles to better navigate through adverse weather conditions and the complexities of urban environments.”
The University of Michigan’s Mobility Transportation Center, now known as Mcity, is a public-private partnership created in 2013 with the goal of leading “the transition to connected and driverless vehicles.” Its home near the heart of Detroit’s auto industry has resulted in a series of partnerships with big names.
University officials said much of the school’s significant work in autonomous vehicles has come in the areas of transportation safety research and large-scale deployments of connected vehicles. In recent years, researchers have put almost 3,000 vehicles with connected technology on the roads in and around Ann Arbor.
That research played a key role in drafting of the voluntary federal guidelines for testing connected vehicles issued last year.
In Palo Alto, many of the major players are nearby — with several, like Ford Motor Co., renting space on Stanford property. But Gerdes said the autonomous projects pursued by the university typically originate in-house, not at the urging of private companies.
And there are plenty of interested companies around. Forty-one industry affiliates have ties to the Center for Automotive Research, while the university has more in-depth partners like Audi AG of Germany, Renault SA of France and Japan’s Toyota Motor Corp.
Tucked inside one of the bays at the Automotive Innovation Facility is an autonomous Ford Fusion nicknamed Trudi — after Gertrude Stein, the American expatriate writer — equipped with sensor arrays and bearing the Stanford logo. It’s part of the university’s partnership with the Dearborn-based automaker.
The building also houses the Volkswagen innovation lab. Some of the work here centers on pushing the performance of driverless cars to the limit, which is where projects like the drifting DeLorean come in.
“The world’s best drivers can harness the full capabilities of the car and we think it’s very possible that automated vehicles can do the same thing,” Gerdes said. “We think it’s very possible that automated vehicles can do the same thing — not because they necessarily want to drift, but because you hit a patch of ice ... the typical driver may over-correct ... or may just have to wait it out as the vehicle slides.
“An automated vehicle should never have to do that. We should never have to worry about the vehicles losing control. We should always be able to figure out what are my options and what paths can I take and make good decisions in that capacity.”
Several bays down from the Ford Fusion sits the X1 — a student-built vehicle test bed for all kinds of technologies. Those include specialized steering and brake controls that allow the car to simulate sliding on ice. At the same time, the setup also allows for the monitoring of driver brainwaves to “understand what humans do in extreme situations.”
Stanford’s research into autonomous vehicles spans departments and disciplines, sometimes reaching into surprising areas. Gerdes says researchers have been involved with backgrounds in aeronautics, computer science and computer engineering, legal, medical and philosophy. That scope is an indication of the far-reaching impacts driverless and connected cars are likely to have in the coming years.
“We don’t try to cover all of the bases,” he said. “What we try to do is identify what we see as key issues that really tie in to faculty passions and student passions. And we work collectively in small teams, often in an very interdisciplinary way, to try to do impactful research along those directions.”
With all of the high-tech research and development underway in Silicon Valley, companies tend to be secretive about their work. Gerdes said Stanford’s autonomous work is purposely made publicly available. That means staying away from contract work with individual companies and utilizing open-gift funding when possible.
“The more openly we do this,” he said, “the faster we can move and the better we can move forward.”