A Self-Driving World

Industry experts discuss what the age of autonomy means for the automotive world

May 31, 2018
by Michael Power

From the April 2018 print edition

For many, it’s a seductive image: you sit scrolling emails, reading the news or staring dreamily out the window as a driverless pod on wheels whisks you around town. This is the image many have of the future of autonomous vehicles, and judging by the media attention that AVs garner, that future appears just around the next turn in the road.

But is fully autonomous driving—so-called level-five autonomy—really that close? As part of its Automotive Intelligence series, which debuted at Toronto’s Canadian International Automotive Show in February, a group of industry experts discussed the topic during a panel entitled Driving in the Age of Autonomy.
Among the panelists was Stratford mayor Dan Mathieson, whose Southwestern Ontario municipality of 31,000 people has become an AV research and testing hub. A demonstration and testing zone has opened there called the Autonomous Vehicle Innovation Network. The centre will see academics and the auto industry working to further AV technology. The response in Stratford has largely been one of excitement, Mathieson said. The region is an industrial and automotive manufacturing hub and there’s been a lot of effort since 2014 to position it as the demonstration-testing zone.

“The residents are getting excited,” he told the audience. “It means we’re going to have state-of-the-art infrastructure. We’re going to be on the cutting edge of 5G and dedicated short-range communication. We’re going to be data collecting off the vehicles. We’re going to be able to start better pinpointing those potholes. We’re going to see our safety standards increase.”

But along with the advantages of becoming an autonomous driving hub come challenges, Mathieson stressed. Cities will have to upgrade their infrastructure and mapping capabilities, and much of the associated technology has a lifecycle of just six to eight years. One question will be whether municipalities can monetize some of that infrastructure. For example, will municipalities be able to earn a royalty on the information that autonomous vehicles are able to collect? “The rubber hits the road in communities and there is where we’ll have our challenges,” Mathieson said.

The University of Waterloo, near Stratford, houses the Waterloo Centre for Automotive Research, and its managing director Ross McKenzie shared the stage with mayor Mathieson. The centre’s focus has been on “perception analysis,” which involves analyzing data that a vehicle collects in real time to help that vehicle make informed decisions. The centre has done work in object detection and enabling autonomous vehicles to differentiate between people and vehicles, McKenzie said. The centre has perfected detecting and determining passenger vehicles, although trucks pose a larger challenge and the focus is on delineating between buses, tractor-trailers, flatbed trucks and others. Another challenge involves pedestrians, he said.

AVs must be able to distinguish between adults and children—or whether someone is pushing a stroller or seated in a wheelchair—as well as other objects that might wind up in front of an autonomous vehicle.
“For autonomous vehicle operation it’s all about not just detecting what’s around you but being able to determine on a highly probable scale, what the likely path for that detected object is,” McKenzie said.

McKenzie saw a few challenges associated with autonomous vehicles. The first challenge, which gets less attention, is the diverse array of technology that’s needed to put autonomous vehicles on the road. A second challenge is dealing with the random nature of vehicles when people are driving them. Autonomous vehicles will have to share the road with human drivers that, for example, drive erratically because they get distracted or try to make up lost commute time by speeding through a yellow light that then turns red, McKenzie said. Additionally, autonomous cars must know where they are going, and high-definition maps for autonomous driving are a key component of that. And while companies are creating these maps, they’re not widely available yet.

“Vehicles aren’t going to go anywhere unless they have these maps,” McKenzie said. “All of the autonomous vehicles that are on the road today have been previously driven by a human operator down the road to learn the road. They don’t go down the road unless they’ve been driven manually first.”

From Pokemon Go to your windshield
Augmented reality is also playing a role in the AV world, said fellow panelist Chris Candy, vice-president of business development at Seven Media. Augmented reality involves laying something digital over the real world, Candy said. Citing the smartphone game Pokemon Go as an example, Candy noted that several companies were using augmented reality to improve their customers’ experience with the brand. IKEA, for example, allows consumers to visualize a piece of furniture in their house. Candy gave a vehicle-related example of an augmented reality repair manual. If a vehicle owner is working on a car but can’t locate a part under the hood he or she would be able to look at a smartphone or table (eventually they will look through glasses) for repair instructions or to identify parts. In the future, windshields may display instructions, such as maps, that drivers currently rely on their smartphones for. “And as everything moves over to the autonomous vehicle side, displaying particular pieces of information (on the windshield) that the vehicle is doing behind their decision process to help people feel a little more comfortable with the autonomous side,” he said.

Fellow panelist Ted Graham, head of open innovation for General Motors, said that the company had been investing in areas where it saw the potential for future growth. And the evolution of autonomy is happening faster than many think, Graham noted. For example, by 2019, the company will have on the market in a major city self-driving “robo-taxis”—Graham described theses as a shared form of mobility used to get around urban environments. GM recently bought San Francisco-based Cruise Automation, a driverless car company founded in 2013. The company is testing Level 4 autonomous vehicles on that city’s streets, Graham said. But GM is also active in Canada’s AV space, he noted. The company recently opened a facility in Markham, Ontario to complement its engineering centre in nearby Oshawa. “I’m here to reinforce that it’s here much faster than you think,” he said.

But when will autonomous vehicles—the driverless, pod-like vehicles many people dream of—actually arrive on our streets? Dennis DesRosiers, president of DesRosiers Automotive Consultants, reinforced that the three major disruptors of autonomous/connected vehicles, electrification and ride sharing are steadily becoming the reality. But the public gets confused regarding the level of autonomy and when it will arrive. And while much of the technology exists for a high degree of vehicle autonomy, other areas such as infrastructure must catch up. It’s also true that most vehicles sit idle 95 percent of the time. “With all of these technologies converging, we’re going to have higher utilization and that’s going to create a lot of efficiencies,” DesRosiers said. “But there are a lot of issues to overcome.”

At the same time, elements of autonomy are appearing in vehicles that OEMs are producing now, DesRosiers said. Active cruise control, automatic parking, blind spot detection and other features represent steps towards autonomy that will spread down to mass-market vehicles quickly. He agreed with fellow panelists that Canada’s public infrastructure must play catch-up to ensure it can support vehicle autonomy—and while it took Stratford eight years to prepare, larger municipalities like Toronto may take much longer. Legal issues must also be considered before full autonomy becomes a reality, DesRosiers said. For example, will autonomous vehicles ever be allowed to break the law? Will all vehicles be compelled to travel 99km/h on major highways regardless of circumstances? Or, will governments establish special AV lanes?

The need for answers to these and other questions will extend the timeline for AV adoption. Full autonomy—the self-driving “pods” on wheels inhabiting the public imagination—is more likely to become a mainstay of Canadian roads in the 2030s or beyond. “All roads lead to it, but I think it’s going to be a longer road than people think it’s going to before the technologies that people are ultimately talking about,” he said. “But the short road is already there with a lot of those technologies.”