MCity and the Race to Get Connected Cars on the Road
The University of Michigan unveiled MCity this week to a throng of international media and local dignitaries on hand to check out the $6.5 million, 32-acre controlled environment specifically designed to test the potential of connected and automated vehicle technologies.
The replica city—complete with movable building facades; roundabouts and traffic signals; bike lanes and a bus stop; tree canopies; robotic pedestrians made to behave just as erratically as their real-world counterparts; gravel and trunk line roads; and a limited-access freeway—is the world’s first full-scale simulated urban environment and outdoor lab for testing the connected car.
MCity also features a control network to collect traffic data wirelessly and from a real-time kinematic positioning system. The test site was built by U-M’s Mobility Transformation Center (MTC) in partnership with the Michigan Department of Transportation, and one of the key goals is to put a shared network of connected, automated (including driverless) vehicles on the road in Ann Arbor by 2021.
MCity is also a powerful opening salvo in the race to see who leads the development of connected and autonomous vehicle technology. Many feel it’s Michigan’s race to lose, as the state is home to the domestic auto industry, 375 automobile technology research centers, and the highest per capita concentration of industrial and mechanical engineers in the nation.
The university has already marshaled an impressive number of researchers, government entities, and private companies in getting MCity on its feet—see the sidebar for details on technologies being tested at MCity by industry partners—but there’s real money to be made from the reams of data connected cars will soon be collecting, and the competition to get in on those profits will no doubt be fierce.
Google is running arguably the best-known autonomous vehicle research program from its headquarters in Mountain View, CA, but there’s still plenty of room for a different leader to emerge. At a separate event showcasing autonomous vehicle technology in Ypsilanti, MI, this week, Google execs defended the safety record of the company’s fleet of self-driving cars after one was involved in an accident on July 1 that resulted in minor injuries to its passengers. More testing is needed, program lead Chris Urmson told reporters, before self-driving cars will be able to anticipate the myriad things that could go wrong in real-world driving.
Other major connected and autonomous vehicle testing efforts underway are Virginia Tech’s Virginia Automated Corridors, 70 miles of roadway in the greater Washington, DC, area, and the GoMentum Station in Concord, CA, a 5,000-acre space at a decommissioned naval base about 30 miles from Silicon Valley that counts Mercedes Benz and Honda as industry partners. A smaller, 10-mile test bed has also been established between Tampa and Brandon, FL, led by the Automated Vehicle Institute at the University of South Florida.
Ed Olson is a computer science professor at U-M who also runs a lab with fellow professor Ryan Eustace that is focused on “robot perception, coordination, and planning.” He’s been working with Ford on a fleet of six, full-size Ford Fusion research cars that scientists are attempting to imbue with human-like driving skills. “It’s very exciting to be a part of MCity,” Olson said at the opening event Monday. A lot of the software work for the Ford research cars was done at U-M, and Olson said he’s looking forward to continuing his groundbreaking work under the MCity umbrella.
Olson is heading up an initiative to create a fleet of 3D-printed, autonomous vehicles to ferry people around campus. Students and faculty would be able to summon the cars with a smartphone app. “Everyone is excited about autonomy, but there are a lot of technical problems with transportation on demand,” he explained.
The idea with U-M’s autonomous fleet is to solve those technical problems—for instance, should the cars have one seat or two; how many cars should be handicap-accessible; how long will users be willing to wait for their self-driving cars to arrive after they’ve been summoned; how should a fleet of electric cars be coordinated when each car’s battery is in a different state of charging; which vehicle model should be sent to each user; where should the cars idle when not in service—using Olson’s low-speed vehicles, and then modifying the campus environment to get the cars into service for further fine-tuning. Then, researchers can apply what they’ve learned from the project to MCity as a whole.
“All of these problems are so cool,” Olson said. “It would be a shame to avoid looking at them until autonomy is solved. We want vehicles we can deploy now so when autonomy is ready, we’ve already thought about the social and economic questions.” And if U-M is successful in developing a self-driving fleet for students to use, Olson envisions many more places where they could be used: corporate campuses, retirement facilities, amusement parks, airports, or city centers.
Local Motors, the Chandler, AZ-based company that made a splash in Detroit earlier this year by 3D-printing a car on the floor of the North American International Auto Show, is partnering with Olson’s lab to develop the fleet of autonomous campus cars. Local Motors custom-built a 3D-printed car for use in Olson’s project. It arrived on campus last week and was on display at the MCity event (pictured below). Over the next year, Olson’s team will develop the vehicle’s self-driving capabilities and build a mobile phone interface for use in requesting a ride.
“We went to Ed, and the project timelines were in alignment with what we wanted to deploy, so it seemed like a perfect fit,” said Corey Clothier, an advanced vehicle technologies specialist for Local Motors in Ann Arbor working on the project. “Local Motors wants to develop a vehicle ready for automation, so we’ll work with Ed’s team to find out what’s needed.”
The vehicle for Olson’s project uses a power train from a traditional golf cart, but the rest of the car has been reformulated specifically for the project and is printed using Lego brick plastic reinforced with fiber to make it tougher. Olson said Local Motors represents the ideal partner because of its ability to print custom components within a matter of hours. “With a 3D-printed car, you can experiment and make adjustments very rapidly,” he said.
Olson said his goal is to demo the first self-driving car in U-M’s fleet next April or May: “Hopefully, people will say this is not a stupid thing and we’ll get funding for 20 or 30 vehicles. Until you try it with data from real-world users, you don’t really know what’s going to happen.”
Olson said his project is a “poster child for MCity” and the various technologies the university is seeking to help develop.
“The challenge is that 98 percent of all driving is incredibly boring and 2 percent isn’t—no two events are the same,” he added. “So many things can go wrong. But we can take thousands of kilometers of real-world driving and, with MCity, we can concentrate that down and we can really make it nasty and mean-spirited just like it sometimes is in the real world. MCity is less a test track and more a robot with sensing, thinking, and acting along with embedded computing. Rather than take the car to the test track, with MCity, we set up the interaction and then the test track puts the vehicle through its paces.”
When Olson first introduced himself to me earlier in our interview, he joked that his lab was doing the same kind of research as Google’s autonomous vehicle program, “only better”—in fact, Olson said some of his former students are now part of Google’s self-driving car research team—and it seems clear that the competition to get these technologies into the market first is something MCity’s constituents are all too aware of.
Olson said, in general, he’s optimistic that a safe, autonomous vehicle is possible, though the journey to putting one on the road might be a long one.
“Building an autonomous car is really hard,” he said. “People travel 100 million miles per fatality in this country, and that level of performance is amazing. We’re going to see new failure models with autonomous cars and it’s not a given they’ll be safer—that’s the challenge. There’s lots of technology still needed. Some of the stuff we can do right away, but some stuff we’ll probably still be scratching our heads about 10 years from now.”
What role Michigan will ultimately play in the development of connected and autonomous car technologies remains to be seen, but Olson agrees that the Great Lakes State has the advantage of proximity.
“Michigan has a grasp on the auto industry, but it would be really easy to take for granted,” he said. “There’s something to admire about the West Coast sentiment of, ‘Screw it, let’s do it.’ MCity is a great down payment on keeping Michigan the center of connected vehicles, but we need to be cognizant that this is a fast-moving field and Detroit is not the only place capable of developing these technologies.”
One reason the University of Michigan’s MCity connected vehicle research center is so unique is the sizeable contingent of prominent industry partners it has recruited. MCity has 15 “leadership circle” companies that have each ponied up money to support its research, including Ford, GM, Nissan, Navistar, and State Farm.
Based in Germany and the world’s third-largest manufacturer of appliances, Bosch is focused on automatic emergency braking technology that combines a radar sensor with a mono video camera to classify objects in a car’s path, enabling a driver warning system that helps prevent or mitigate front collisions with pedestrians.
Delphi is one of very few auto suppliers worldwide capable of providing and integrating all of the technologies required to make autonomous vehicles a reality. Delphi has already developed an automated vehicle that has traveled cross-country, and the newest generation of its driver alertness-sensing technology, delivered through a tablet interface, helps reduce driver distraction.
Denso, a major world supplier of automotive components and technology, is developing tools to make autonomous technologies work together, including dedicated short-range and data communication modules, driver status monitoring, and head-up displays.
A leader in transportation management tools since 1933, Econolite is developing a vehicle-to-infrastructure kiosk that includes traffic-controlling software and other connected vehicle communications that uses radio frequency technology to facilitate safer driving.
The global automaker demonstrated the potential of dedicated short-range communications between vehicles and smartphones through a proprietary vehicle-to-pedestrian app that can alert both driver and pedestrian to an imminent collision.
An IT company concentrating on the transportation and agriculture industries, California-based Iteris is focused on predictive traffic and weather analytics, reducing congestion, and improving safety. Iteris will integrate data generated by bicycles and vehicles at MCity into its research.
Specializing in mobile and wireless technology, Qualcomm’s Vehicle Integration Property car was on hand to demonstrate three of its top automotive advances: Halo and WiPower wireless charging technology, as well as dedicated short-range communications.
The world’s largest auto manufacturer exhibited its Vulnerable Road User test mannequins—adult, child, and bicyclist versions—used in standardized safety testing.
The telecommunications giant is seeking to advance car-sharing technology that runs as an app on any smartphone or tablet. The company’s technology is currently being used in the first large-scale, pay-by-the-mile road usage program in the U.S., happening now in Oregon.
The document management and business process corporation is developing video and short-range technologies capable of detecting and transmitting information about the number of passengers in a vehicle for use in determining a vehicle’s eligibility for high-occupancy vehicle lanes, as well as connected parking technology that guides drivers to open spots.