All anyone in the automotive industry seems to be talking about these days is the race to develop autonomous cars and new mobility applications. The announcement last month of a new $1 billion Toyota vehicle research operation in Ann Arbor, MI, is another significant milestone in Michigan’s quest to lead the nation in the development of these technologies.
The Toyota Research Institute (TRI), in partnership with the University of Michigan, will be Toyota’s third autonomous vehicle research center in the U.S.; the other two are in Palo Alto, CA (with Stanford University), and Cambridge, MA (with MIT). The TRI is scheduled to open next month and will eventually employ 50 people. The institute will fund research in artificial intelligence and robotics to support fully autonomous driving and to accelerate discoveries in materials science.
“Our goal at the TRI is to make the safest, best vehicle in the world that is both uncrashable and convenient to drivers,” said Ed Olson, an associate professor in electrical engineering and computer science who will lead research at the institute along with Ryan Eustice. MCity, the 32-acre replica town that opened last summer on the northern edge of U-M’s campus to test the potential of connected and automated vehicle technologies, enhances the work Toyota is doing, he said.
At the TRI, Olson will study how robots perceive and understand their environment—research often involving LIDAR and sensors. Eustice will study mapping and localization. Toyota is already an industry partner in the Mobility Transformation Center, a U-M-based collaboration between industry, government, and academia that oversees the work at MCity, and is laying the foundation for a commercially viable system of connected and automated vehicles. Ultimately, U-M plans to deploy vehicles in Ann Arbor that provide an on-demand mobility service.
Olson said Toyota was attracted to Ann Arbor for a number of reasons: the company already has two technical centers located in the Ann Arbor area, it has a strong relationship with the university, and it wants to be close to the American Center for Mobility, set to open sometime in the future at the former Willow Run Airport site just outside of town.
“What makes Toyota different from Google, Ford, and other [autonomous vehicle] players is the longer view they take of the technology road map and their partnership with universities on basic research,” Olson said. “Testing is one of the biggest challenges we have with autonomous vehicles. It’s not hard to make a YouTube video, but how do you make the car as reliable as humans, who go 100 million miles per fatality? No car has ever been tested 100 million miles before release.”
Olson said the plan is a slow, deliberate testing process where researchers intentionally create “difficult, nasty cases” for automated cars to successfully navigate. “The goal is not to build one automobile, but to create autonomous technology that can be put in every car,” he added. “We’re not in startup mode—we’re not trying to make a cool prototype and then sell it to an OEM for $1 billion. We’re in this for the long haul.”
Olson said the TRI’s mandate from Toyota goes beyond autonomous vehicles, and described his upcoming work as “robotics writ large.” The institute aims to enhance the safety of automobiles; increase access to cars for those who otherwise cannot drive; and accelerate scientific discovery by applying techniques from artificial intelligence and machine learning to other areas in order to lower costs and improve performance. Fifteen employees from Toyota’s local tech centers will transfer to the TRI when it opens to continue ongoing research.
As for how Michigan is faring compared to other national innovation hubs in the race to develop driverless cars, Olson said the work already being done at MCity, the Mobility Transformation Center, and elsewhere in the state to develop connected vehicle technology is creating a “positive feedback loop” and attracting some of the best and brightest students to the field.
“Michigan is in a great position given its existing [auto manufacturers] and academic environment,” he noted. “Michigan also has a good regulatory environment compared to California, or really any other state. We’re well positioned to lead the development of autonomous vehicles, but we have to keep working at it. It’s not a foregone conclusion.”
