Skip to main content
Will letting a robot do the driving mean we program our way to error-free transportation?

At the University of Waterloo, the self-driving Autonomoose vehicle is already a reality. During the initial series of testing, this Lincoln MKZ based-model has a human in the driver seat just to keep an eye on things in case something doesn’t work out as planned.

But Autonomoose is expected to move quickly from this Phase 1 testing to Phase 3 when the human in the vehicle can safely turn his attention to something besides the road. Ultimately, Autonomoose will reach Phase 4 when its passengers or cargo can ride safely without any human intervention.

Welcome to the world of safe autonomous transportation. Self-driving vehicles are a reality – not a Saturday morning cartoon. Researchers at the University of Waterloo are developing safety features into the software that allows self-driving vehicles to detect and respond to other vehicles, stop signs and traffic lights. The vehicles will also have radar, sonar, vision sensors and enough computer power to efficiently use the data they collect.

The serious safety issues presented by these cars aren’t actually in the technology itself, researchers say. The biggest problems are related to road and highway infrastructure and mapping. The driverless car has to know where to go, so it needs a highly accurate map and a predictable roadway. The U.S. National Highway Transportation Administration released an automated vehicles policy in September 2016, which aimed to move state traffic safety administrations in the same direction so there would be consistency in road modification throughout the United States. These polices are likely to influence Canada as well.

Safety is one of the biggest selling points of self-driving cars, says Barrie Kirk, executive director of the Ottawa-based Canadian Automated Vehicles Centre of Excellence, or CAVCOE, which is a nonprofit dedicated to helping the public and private sectors prepare and plan for the arrival of automated vehicles.

The biggest risk to riding in conventional vehicles, Kirk says, is the possibility that the driver – or some other driver – will make a bad decision: go too fast, too slow, weave, fail to pay attention, drink and drive or fall asleep driving for example. Computers lack these human foibles. That’s why in the long run, self-driving cars have the potential to be a much safer alternative, Kirk points out.

In Canada alone, auto accidents are a major cause of death. “Traffic collisions currently kill about 2,000 people in Canada annually,” Kirk said in a report prepared last year for the Government of Canada. “We hope and we expect that autonomous vehicles will be able to prevent more than 80 percent of all road collisions, injuries and fatalities.”

Autonomous vehicles will have the further advantage of encouraging pedestrians to have priority, especially during certain times of the day, Kirk says, which will make them safer. Because people feel safer, they may be more likely to walk and, “walking promotes health and well-being,” Kirk’s report says.

Kirk also points out that an estimated 80 percent of resources in a hospital emergency room are devoted to caring for victims of automobile collisions. “If we can prevent 80 percent of these collisions that will mean a greatly reduced workload for hospitals and emergency divisions,” he says.

Some skeptics have raised the fear that autonomous vehicles could be vulnerable to hackers, who might use the hack to put passengers in danger. Morgan Stanley Research looked at this potential problem and concluded that hacking a car wirelessly is difficult, but not impossible. But instances of “car hacks” have so far been physical — wires connected from the hackers’ computer to a car’s onboard diagnostic system while the “hacker” was inside the car. “The risk in this situation is the same as the risk that a burglar is sitting in the back seat with a gun to your head,” the Morgan Stanley report concluded.

Nevertheless, setting up ways to prevent this from happening is one of the challenges developers of autonomous vehicles are addressing, Morgan Stanley said.

Morgan Stanley actually sees a bigger safety risk coming during the decade or two that it will take to retire most people-driven vehicles. With more than 1 billion conventional vehicles on the roads worldwide, these researchers warn that getting people to accept and buy autonomous cars is a big problem. They warn of a witch’s brew of autonomous and conventional vehicles unhappily sharing the roads and dramatically increasing the number of accidents.

One solution they suggest that might make even the most resistant customer for autonomous transport happier is for someone – the government or automotive manufacturers – to pay people to scrap their conventional vehicles in favor of autonomous ones. This could “cut the time needed to achieve full [acceptance] by half,” Morgan Stanley predicts.

That sounds like a safe bet.