Self-Driving Car Makers Broaden Their Test Fleets

Industry leaders Waymo and GM graduate to larger, public test programs.

The drive toward autonomous vehicle technology gained momentum last week as two major players laid plans to put more test vehicles on the road.

Waymo LLC , formerly known as the Google self-driving car project, invited residents of Phoenix, AZ, to be part of its “early rider program,” which calls for hundreds of self-driving vehicles to be made available to families and commuters. Meanwhile, General Motors Co. reportedly filed with the Federal Communications Commission to add 300 more self-driving cars to its existing test fleet of 50 autonomous Chevy Bolts, according to a report in the Detroit News .

The actions of the two companies are a small step forward for the autonomous car segment, which to date has mostly busied itself with announcements, investments, acquisitions and secretive engineering work.

In a statement on its website, Waymo said the goal of its new program is “to give participants access to our fleet every day, at any time, to go anywhere within an area that’s about twice the size of San Francisco.” The program is the first public trial for the company since it began working on self-driving technology in 2009.


Waymo (formerly known as the Google self-driving car project) said that it is adding 500 self-driving Chrysler Pacifica minivans to its fleet, bringing its total to about 600. The minivans will be publicly tested in Phoenix, AZ. (Source: Waymo LLC)


Waymo said that it is also adding 500 self-driving Chrysler Pacifica minivans to its fleet, bringing its total to about 600. The company announced in December that it was partnering with Fiat Chrysler Automobiles (FCA) to outfit 100 hybrid Pacificas with its autonomous driving technology. Under the partnership, FCA engineers integrate Waymo’s technology into the electrical architecture of the Pacificas.

GM announced publicly in December that it was testing self-driving Chevy Bolts on public roads in Michigan, but last week it would not say whether its fleet is growing. “We have GM automation engineers testing more than 50 vehicles, but we aren’t commenting on the specifics of a broader fleet,” GM spokesman Patrick Sullivan told Design News last week.


In December, GM CEO Mary Barra announced that GM would begin testing self-driving Chevy Bolts on public roads in Michigan. (Source: General Motors Co.)


To be sure, the announcements don’t mean that consumers will soon be enjoying curb-to-curb service from self-driving vehicles with no drivers. In its statement, Waymo said its autonomous Pacificas will have drivers aboard. “In the initial stage of the early rider program, we’ll have a test driver at the wheel, allowing us to gather more feedback, develop self-driving features and refine our technology,” Waymo wrote.

Industry experts said the Waymo program is a “limited domain” test performed under well-understood conditions. “It’s a large geographic area, but if you look at the conditions, it’s fairly consistent across the area,” noted Sam Abuelsamid, a research analyst for Navigant Research . “It’s an area with warm weather all the time, and it doesn’t get a lot of rain.”

Despite claims by automakers that they will release full, self-diving


Was it the intention of Google to make Waymo look like the Ghostbuster's station wagon? If so, they nailed it.

We must ban all autonomous experiments immediately. They are insane. Humans do vision recognition at least a hundred million times faster than computers because the human brain has hundreds of billions of parallel processors. These autonomous experiments can't even recognize turn signals or brake lights, or differentiate between a bus and a car. They are incredibly dangerous, inherently. They have to rely on GPS mapping, and we all know how out of date that can be.

This is going to awesome and we will certainly see this happen. hello from india i am reena from covering all cars news <a href="">Carwaar</a>

Never going to happen. Rules of the road are far too subtle to ever be captured in a decision tree that computers are way too slow to evaluate anyway. Plus would you really want terrorist to be able to kill at will without risk?

Kirk, I agree with you that attempting to duplicate human visual recognition is a losing proposition. I think there are other considerations, however: 1) humans don't come close to using their full capacity in order to drive, and 2) automation can provide other (possibly superior) means of detection (radar, infrared, Wi-Fi communication with other vehicles, 360 vision, etc.) We'll see how it goes.

I do agree that we could provide more information to assist a human driver. But I don't like radar because if everyone has a half dozen radar projectors, people will end up getting cancer. Communications would be much better than things like turn signals or brake lights we rely on now, but has the disadvantages of ending privacy of where you drove, as well as inviting hacking. With lidar and infrared, we have to consider how reliable these systems will be after 10 years or so?

Add new comment

By submitting this form, you accept the Mollom privacy policy.