Riding in a self-driving car will change how you think about the future of transportation.
When I rode in one of Google’s Lexus self-driving cars last fall, I was riveted by the screen above the dashboard that displayed the cars and other major objects around us. We could view their every move in real time.
In other words, the car had a lot of knowledge about its surroundings—way more than any human driver.
But twice, the Google software engineer who was demonstrating the car for me had to take control of the wheel due to bugs that prompted it to shut off its autonomous-driving mode. He also said the self-driving mode didn’t yet work in bad weather like snow and fog.
The technology clearly has a long way to go, and Google has said it would be ready within five years.
Many questions remain about whether self-driving vehicles will be able to drive more safely than human drivers. In the U.S., about 10 million car accidents occur every year, according to federal government data. There is roughly one fatality for every 100 million miles driven in the U.S., or more than 33,000 deaths a year. Fatality rates generally are much higher in less-developed countries.
Self-driving cars, of course, haven’t logged that many miles. Google’s fleet of retrofitted Toyota cars have traveled more than 500,000 miles without causing an accident, the company has said.
One person familiar with Google’s self-driving project said getting the software to be perfect remains a big challenge.
“99% isn’t good enough,” this person said, noting technology would have to be even safer than that in order to win over the public.
There are other skeptics. Elon Musk, CEO of small luxury car maker Tesla Motors, earlier this year said he thought Google’s self-driving car system was too expensive.
Daimler, General Motors and other car companies offer some self-driving features as add-on packages for luxury vehicles. Robert Bosch, the German electronics and auto-component supplier, has said it hopes to commercialize a “traffic jam assistant” system that can “brake, accelerate, and steer completely autonomously” in traffic. It could sell the system to major car brands.
Google has been talking to automakers to try to license its self-driving technology into their new cars, and it has retrofitted Prius and Lexus cars to show how it can be done crudely.
But Google also has bigger plans and has been studying how its vehicles could become part of robo-taxi systems in which a fleet of cars would operate autonomously in a particular area, picking up passengers and work commuters on demand, according to people familiar with the matter.
Google believes such robo taxis could reduce the need to own a car. Some existing car-sharing services, such as ZipCar and RelayRides, have said that each shared vehicle can eliminate the need for 10 privately owned vehicles. Some industry observers have been skeptical of such estimates.
Google and other self-driving car proponents believe that “robo cars” will eventually be better drivers than humans, reducing fatality and accident rates. They also argue that such cars could be lighter weight with better gas mileage. Another theory goes that if most vehicles on the road drive autonomously, proponents believe there will be fewer traffic jams and cars could move faster.
“Today’s car purchaser asks, ‘What car do I need for my life?’ and then they buy a big SUV to take them skiing twice a year,” says Brad Templeton, a well-known software engineer and entrepreneur who advised Google’s self-driving car project in the past. In the future, he said, people instead will be able to ask, “What car do I need today?” (He declined to comment on Google’s efforts.)
That would upend the economics of automobiles. More cars would be manufactured but they would get more usage and need to be replaced more frequently than privately owned cars.
It could also upend the economics of insurance. If you don’t own a car, you may not need car insurance. In thinking about the robo-taxi model, Google has considered the idea that it would be responsible for insurance, according to a person familiar with the matter.