3 Myths about Self-Driving Cars

There’s a lot of hype surrounding them. But where are they?

Driverless cars seem poised to be the next big thing in the auto industry, or maybe the next big thing after that. Google (GOOG) routinely turns heads on California roads with its famous driverless cars, which have logged half-a-million miles with barely an incident. California, Nevada and Florida have passed special laws allowing self-driving cars, with more states sure to follow. The federal agency that regulates car safety recently issued new guidelines for automated vehicles and might even begin to require some driverless technology that could significantly improve safety.

Piper Jaffray (PJC) analyst Gene Munger estimates that driverless car technology could ultimately mushroom into a $240 billion industry, which would be roughly three times the size of the cloud-computing industry that will supposedly transform data storage. Yet there’s a lot of confusion about what self-driving cars are, who’s likely to build them, and how the new technology will transform driving. So here’s a quick guide to the three biggest myths about self-driving cars:

Myth 1: They’re far off in the future. Gee-whiz stories about self-driving cars make it sound as if a fleet of robotic pods will materialize at some distant date and revolutionize driving overnight. In reality, driverless technology is already migrating slowly into new vehicles through features such as automated braking, which applies the brakes if you get too close to a car ahead of you; adaptive cruise control, which speeds or slows the car based on the speed of surrounding traffic; self-parking systems that use cameras to help back a car into a parking space; and lane-departure warning systems that read lane markings and sound an alert if your car drifts beyond them. Like a lot of new automotive technology, these types of systems tend to be pricey and available mostly on luxury makes. But they'll get cheaper and more commonplace.

Cars will become driverless in stages, as new software begins to tie together various automated systems. Steffen Linkenbach, an engineer who oversees North American automated vehicle programs for supplier Continental Automotive, foresees three basic phases. By 2016, he says, partially automated cars will be able to brake and steer at speeds up to 20 or 25 miles per hour, which ought to help reduce rear-end crashes in particular. By 2020, there will be highly automated cars that do the same thing at highway speeds.

You still won’t be able to recline the seat and take a nap while the car shuttles you around, however, because the driver will have to be ready to take over on a few seconds’ notice in case there’s a problem. “In reality, there’s rain, sleet, snow, and sometimes a deer jumps out in front of you,” says Paul Perrone, a roboticist who’s chairman of a committee at the Society of Automotive Engineers that sets standards for automated vehicles. “When it happens, you have to deal with it. You can’t be asleep hoping the technology will handle it."

What drivers might be able to do, however, is take their eyes off the road momentarily to send email on a smartphone or do an Internet search. “It won’t be necessary to have 100 percent of your attention on the road,” says Linkenbach of Continental.

Fully automated cars that do all the driving all the time are probably much further off. Technology isn’t yet able to handle complex tasks such as deciding who should go first when several cars stop simultaneously at a four-way intersection, or construction blocks part of the road, requiring you to veer into an oncoming lane to get around it. Fully automated driving may also require a network of sensors in roads and other cars, plus new laws that account for liability when something goes wrong. Given the fractious nature of American politics, fully driverless cars might be more likely to originate in someplace more orderly such as Sweden (where Volvo is based) or Finland.

Myth 2: Self-driving cars will be unsafe. Many people like to feel in control when they’re mobile, which tends to generate an instinctive sense of safety. But there’s convincing evidence that humans overestimate their ability as drivers, and that computers can handle many elements of driving far more safely than humans. For starters, more than 90% of crashes are caused by driver error, so putting computers in charge of more things computers are good at should cut down on dangerous driver mistakes.

“The hope is that turning more mundane things over to automation would result in an environment where people’s minds wandering doesn’t lead to tragic results,” says David Zuby, chief research officer for the Insurance Institute for Highway Safety, a research group backed by insurance firms.

Some automated safety features have been around long enough for researchers to gather real-world data showing their effectiveness. Electronic stability control — which takes over steering and braking in a skid, to help the driver regain control — reduces the risk of a fatal, single-car crash by nearly 50%, according to IIHS. That’s a huge impact for a single safety system. IIHS says Volvo’s CitySafety automated braking system has led to a 27% reduction in insurance claims for car damage on autos equipped with it, compared to similar vehicles without the system. Other technologies show similar improvements in safety, which is one reason the U.S. driver fatality rate has fallen during recent years and is close to record lows.

Assigning computers a greater share of the driving burden will increase the chances of a technical malfunction that causes a crash and makes drivers jumpy about trusting machines. That’s why automakers are likely to build many redundancies into automated cars and make sure there’s a foolproof way to signal when the driver needs to take control.

Myth 3: Google will be the leading purveyor of self-driving cars. Google has garnered a lot of press attention with its fleet of vehicles topped with contraptions that look like a giant helmet-cam (it’s actually a $70,000 laser-based rangefinder system). But Google doesn’t build cars, and the company has been cagey about what exactly it plans to do with the technology other than “to improve people’s lives by making driving safer, more enjoyable, and more efficient,” according to a spokesman. Instead of building cars or car components, it’s possible Google is more interested in precision mapping, which is necessary for fully automated cars and is one of Google’s core competencies.

Meanwhile, automakers such as Audi, Mercedes, Cadillac, Lexus and others are developing automated cars that are likely to start appearing in showrooms within the next few years. And third-party firms such as Mobileye are developing off-the-shelf driverless technology for a fraction of the cost of Google’s $150,000 system.

Self-driving cars are definitely coming. They just might look more familiar and less exotic than expected.

Rick Newman’s latest book is Rebounders: How Winners Pivot From Setback To Success. Follow him on Twitter: @rickjnewman.