Connecting state and local government leaders
“Just the fact that there are number of AV microbus experiments already up and running suggests there is indeed a problem to be solved,” according to tech analyst Mark Rogowsky.
ANN ARBOR, Mich. — Amid the flurry of hype about an imminent national rollout of fully autonomous vehicles, one vehicle type may actually be ready to hit the streets: the AV microbus.
This week, MayMobility, a local startup, is trialling a fixed-route AV microbus shuttle service for employees of Bedrock Detroit, the commercial real estate firm of entrepreneur Dan GIlbert. Travelling on a “simple, three-quarter mile loop between a parking garage and an office building,” according to MayMobility’s CEO, Edwin Olson, the six-passenger microbus may confirm what some industry analysts have long suspected—that the reality of autonomous vehicles may come not in a flash, but in baby-steps.
“One of the cool things about putting AV on fixed routes is that some of the most difficult, unsolved problems start to go away,” according to Mark Rogowsky, former communications head at Uber, and now a Seattle-based technology analyst.
Tesla, GM, Google’s Waymo, Uber and an array of smaller companies are of course in the midst of a technology race to deploy fully autonomous, personal vehicles. But the claims made in their corporate press releases have often run well ahead of actual, successful trials. Why might that be?
Landscapes in the built-environment are often highly varied, for example. The transition from highways to boulevards, and then to urban centers and residential neighborhoods, have proven, so far, to have overwhelmed AV trying to navigate such information-rich topography. Despite very advanced mapping capability, these broader AV tests have often been confounded by unexpected pedestrian behavior, road construction, or sudden decays in street conditions. AV at first seem highly capable, until they don’t.
“Recent reports about Waymo vehicles having trouble making a left turn puts a fine point on how even simple things may be, at random times, surprisingly difficult,” Rogowsky said in an interview.
Extensive route pre-testing was performed in Detroit before MayMobility rolled out its microbus, with engineers walking the targeted route. “While there’s a certain amount of wrinkle and complexity we can handle, we do like features such as dedicated bike lanes, protected turns, and two-lane boulevards that run in both directions,” Olson said. “We don’t think it’s technologically possible, right now for example, to sell a customer a personal AV and have that AV still running perfectly six months after purchase. We, on the other hand, are in a wildly different segment of this business.”
MayMobility is one of several companies globally with AV microbus trials already underway. Navya, a French mobility company, has already tested vehicles in the Parisian business district La Defense, and conducted trials in cities as diverse as Las Vegas and Christchurch in New Zealand.
Seeing a growth opportunity, Navya has arranged to put AV microbuses into production in Saline, a city of 9,100 residents about 10 miles south of Ann Arbor and will also be deploying a vehicle on the campus of the University of Michigan later this fall.
Indeed, Washtenaw County, which includes Ann Arbor, has become a new center of AV and mobility research led by the university’s M-City AV testing facility, a walled garden of streets and village facades that looks like a Hollywood backlot.
In September, Route Fifty had a chance last month to walk the proposed route of the Navya-built AV microbus, between the university’s Lurie Engineering Building and North Campus Research Complex.
The course runs over lightly trafficked North Campus roads that are smooth, turn gently, and are free of unexpected hazards. Speaking casually with students, they were quite aware of the coming launch of the microbus and said they’d be eager to take it for a ride.
AV microbuses are not high-speed vehicles. The MayMobility offering is built on top of a standardized EV chassis produced by Polaris, and the maximum speed of these microbuses is typically 25 miles per hour. That in itself reduces a lot of risk, but also may pave a faster pathway for their adoption.
True, as in broader EV testing, MayMobility will have a representative on board the vehicle to take control if necessary as its microbus traverses its downtown loop in Detroit. But in another sign of how AV microbuses may be closer to actual deployment, the standard sizing of this vehicle type offers an advantage. “We have additional customers lining up now for our AV solution and, you see, many of them have already been using the Polaris GEM chassis which has an existing market in [non-AV] shuttles,” Olson said.
“Much of the AV experimentation today involves retrofitting existing vehicles,” Rogowsky said. “So it’s easier, perhaps, to make a stack on top of an adaptive vehicle. It’s telling that the first signs of standardization are appearing in this AV microbus segment.”
Indeed, as you look at a company like MayMobility, that seems to be the approach. Olson, the company’s CEO, is also an associate professor of computer science at the University of Michigan and director of the Autonomy, Perception, Robotics, Interfaces, and Learning lab, better known as APRIL. And Alisyn Malek, the COO, was formerly with General Motors.
The company’s process is to take an off-the-shelf chassis, and add (stack) the necessary AV layer on top. While other players are trying to solve the entirety of the future AV problem in a single grand sweep, the AV microbus space seems to be aiming for a smaller, but interestingly, a more practical and manageable target.
“Just the fact that there are number of AV microbus experiments already up and running suggests there is indeed a problem to be solved. But it also suggests the technology is already sufficiently diffuse to allow for ready experimentation,” Rogowsky said.
A spirited conversation taking place today in AV technology concentrates on future winners and losers, and whether owning the software, or the hardware, or both (or neither) will ultimately offer the best business advantage. As Olson pointed out, a typical consumer passenger car is set to run for about 30,000 trouble-free miles. That’s a meaningful point. Because, when you consider the crowded effort to get AV into this single-car market, is that really where AV will find their first adoption?
Navya, for example, is marketing their AV microbus to hospitals, airports, industrial parks, amusement parks, and resorts. Voyage, another AV company, has started testing in an enclosed retirement community in San José. Roughly speaking, not until AV have solved the entirety of the navigation problem can the technology realistically take on the single-car market, either for consumers or fleets of ride-sharing vehicles for hire.
“Right now,” Olson said, “AV on the street are a high touch operation, and require constant care and feeding.”
Route Fifty also had a chance to visit MayMobility’s AV microbus onsite at the Technology in Motion event in downtown Detroit in early September. The vehicle was roomy, with lots of light, and had a solid build. The front of the vehicle’s interior boasted a large, convex display screen and it was easy to imagine stand-up passengers moving easily in, and out, through the large retractable side doors.
Malek happily pointed out “our vehicle will be the first AV to test actually test on the streets of Detroit.” Given the notable recovery taking place in downtown Detroit, as the spine of Woodward Avenue is steadily refurbished with a streetcar and revived buildings, the city suddenly seems the perfect place to try out a new approach to urban solutions.
Gregor Macdonald is a journalist based in Portland, Oregon and has written for Nature, Talking Points Memo and The Petroleum Economist.