Google's driverless car initiative is the most public program for developing self-driving vehicles. Here's how far the technology has come and what it still needs to do.
Driverless cars are among the few monumental innovations that represent the massive changes in daily life that will be brought about by new technological advancements. Transportation as we know it is changing, and computers are leading the charge.
The Google Self-Driving Car project is, perhaps, the most public-facing initiative in the space. Google operates a fleet of cars, predominantly Lexus SUVs, equipped with a LIDAR system that uses a laser systematically map and remap its environment as it moves. The cars also use cameras and radar technology to monitor space and other drivers.
A working prototype of a car built by Roush especially for Google was unveiled in 2014 and Google recently began calling for artist submissions for the vehicle's design. So far, Google's autonomous fleet has logged more than one million miles on public roads in a few states in the US, but the prototypes have only driven in Mountain View, California.
"We have driven our Prius and Lexus vehicles across California, as well as in Nevada, Washington, D.C, Arizona and Florida," a Google spokesperson said. "We've driven on a wide variety of terrain through a range of road conditions, including a loop around Lake Tahoe, a trip to and through LA, across all the Bay Area bridges, down Lombard Street, and even through the Santa Cruz Mountains."
The journey to the Google Self-Driving Car began with the DARPA Grand Challenge in the early 2000s for autonomous vehicles, the official term for self-driving or driverless cars. Universities such as Carnegie Mellon and Stanford competed for top honors, with Stanford winning the first year.
Google began recruiting some of the top researchers from the two universities and other competitors. Sebastian Thrun, who was with Stanford at the time, was brought on by Google to work on their Street View project and eventually headed up the Self-Driving Car project.
According to Gartner's Ken Dulaney, the early DARPA contests weren't very impressive. Contestants were not able to fully navigate a short course in Nevada. Contrast that to today, Dulaney said, where Google's latest Self-Driving Car monthly report shows only 12 accidents over the life of the project, with no fault to the autonomous vehicle in any of the accidents.
"I think we can conclude we have made tremendous progress," Dulaney said. "I expect many, many cars to have some of these features in the next two years and they will dramatically reduce accidents. Consensus of the manufacturers are that safety is a huge buying motivator today."
Dulaney said he expects we will indeed see at least some commercial self-driving cars by the end of the current decade and potential mass adoption beginning in the 2020s.
So, it's apparent that the technology as a whole has made tremendous strides and Google's project is pushing the concept along. But that still doesn't answer the questions around why Google is working on this or what it will do with the technology in the long run.
"My sense is that Google's focus has been on developing the technology to make it possible," said Forrester's Frank Gillett. "I'm not clear that they actually want to go into the business of building cars."
By building out prototypes and dedicating so many resources to the development of the autonomous car idea, Google gains insight into technical issues and gets a clear look at the potential business opportunities that come with it, Gillett said.
At its core, Google is a company that survives — and thrives — on the collection and leveraging of data. In a way, their Self-Driving Car is just another avenue by which Google can collect more relevant data to power their foundational products in search and advertising.
As the system constantly checks and maps the world around it, Gillett said he could see Google licensing the technology to a partner company as long as the partner agreed to send information collected by their vehicles. Information collected could tell Google when it needs to go take new photographs for Street View or update some of the traffic patterns in Maps.
The Self-Driving Car technology has the potential to gather what Gillett called the "macro schedules" of people. So, it could potentially gather better information on the movement of people through spaces. In addition to collecting data, it could broaden Google's core business as well.
"I've got to think at least some of these business initiatives, such as self-driving cars, are part of exploring the potential for other fundamental sources of revenue besides advertising," Gillett said.
Still, challenges remain, starting with governmental regulations. So far, only a handful of states have addressed autonomous vehicles in legislation. According to Stanford, state bills allowing autonomous vehicles have passed in four states: California, Nevada, Michigan, and Florida. Bills have failed in nine states and bills are under consideration in 15 states and Washington DC.
If autonomous vehicles gain traction, there will inevitably be a period of mixing, Gillett said. As driverless cars begin to share the road with traditional automobiles, additional rules and regulations may be needed to determine how the vehicles will be allowed to interact or if certain roads will be designated for a specific type of vehicle.
The biggest challenge, however, will be a cultural one. Gillett said that getting the people to see their cars as an appliance instead of a symbol of freedom is the biggest hurdle facing Google and other companies pursuing autonomous vehicles. The automobile is a symbol of so much of a person's life that letting go of some of the independence associated with it may be difficult.
- Google's biggest 'moonshots' in photos
- How big of a threat is Facebook's ad play to Google?
- How Google's gesture control technology could revolutionize the way we use devices
- Google solidifies its IoT strategy with its Brillo OS