UVic's Lin Cai is helping cars talk to each other
More vehicles driven by algorithms means a safer, more sustainable transportation system. Just don’t expect Jetsons-style automation anytime soon
This fall, we’re promised, ride-hailing will arrive in B.C. after many years of obstruction and obfuscation. In fact, you may even be reading this column right now from your mobile phone, seated in the back of a Lyft or Uber vehicle.
Ride-hailing is a much-needed piece in our transportation puzzle. But for the two big companies involved—each of which is losing hundreds of millions of dollars (Lyft), if not billions (Uber), annually—the end game is autonomous-drive vehicles. Get rid of your No. 1 cost—human drivers—and suddenly the ride-hailing bet starts paying off. Uber, for one, is all in, spending upward of US$20 million a month on developing self-driving technologies.
For those researching how to bring more smarts to our roads, autonomous vehicles are only part of the puzzle—albeit the flashiest piece. “In a controlled environment, with everything being autonomous-drive vehicles, we can pretty much ensure a high level of safety,” says Lin Cai, an engineering professor at UVic, who in May was awarded an E.W.R. Steacie Memorial Fellowship by the governor general for her work on wireless networking solutions. “The trouble is when an autonomous driver drives a vehicle on a road mixed with human drivers.”
Cai’s field of specialty, vehicle-to-everything networking, or V2X, takes the concept of the Internet of Things and applies it to our transportation ecosystem. As she sees it, human drivers can learn the communication protocols used by autonomous vehicles, but “we probably cannot ensure all human beings will follow those protocols strictly.”
When protocols are ignored, or sensors fail, tragedy ensues: Uber suffered a public relations nightmare last year when one of its autonomous vehicles killed a pedestrian in Arizona, setting its program back years. For now, then, autonomous vehicles are stuck on the unregulated roads of industrial parks and gated communities (see below).
While companies from Uber to Google work out the kinks in the technology—and politicians and transportation planners debate the practical considerations of a driverless future—Cai and her colleagues are working on solutions for the here and now. Most of the engineering focus in the past 10 years, she notes, has been on wireless tech that allows for human-to-human communication. Increasingly, the need will be for machines to communicate with each other—including car-to-car, cellphone-to-car, and between other roadside sensing devices (such as cellular towers) and the people (or, eventually, machines) on the road.
“Our modern transportation has lots of big challenges, including safety and energy issues,” Cai says. “Once vehicles can talk to each other, we can address many of these challenges.” Autonomous vehicles will demand a high-speed network, but our current system—with all its human foibles—can benefit from more connectivity, too. If we know more about the speed, location, destination and travel conditions for every actor on the road, at any given time, we can better achieve certain outcomes: optimal speeds that reduce carbon emissions, for instance, or dynamic signalling that limits congestion and accidents.
Getting more connected isn’t as easy as flipping a switch, though. “Right now, it’s very expensive to get the licence for wireless spectrum,” explains Cai. “And in the future, with the Internet of Things and V2X, it will become even more expensive.” There are also issues of privacy and system security. And while it may be many years, if not decades, before autonomous vehicles replace human-driven ones, there’s little doubt that the era of the single-occupancy car is firmly in the rear-view mirror.
“Currently when I buy a car, I probably only use it 5 percent of the time,” Cai says. “When we have autonomous driving—plus this wireless communication network—we can schedule cars to arrive at the location where I need it. It will substantially change the utility of vehicles.”
As the world waits for self-driving cars to hit our roads, Optimus Ride is bringing the technology to select controlled (or what it calls geo-fenced) spaces across North America. The Boston-based startup, launched in 2015 by five graduates and researchers from MIT, claims to have provided more than 20,000 autonomous vehicle rides so far.
Besides Boston’s Seaport neighbourhood, Optimus serves Brookfield Properties’ Halley Rise development in Virginia, Paradise Valley Estates in California and, as of August, the Brooklyn Navy Yard in New York City: an industrial park with some 400 manufacturing businesses and 10,000 employees, where a total of six Optimus vehicles are expected to transport 500 passengers each day.