Google Readies its Self-Driving Cars for Real-World Testing On the Roads of California

Just under a year ago, Silicon Valley software giant Google announced the next stage of its ongoing autonomous vehicle research program: to design and build 100 pod-shaped fully-autonomous prototype low speed electric cars that would, in its own words, “shoulder the entire burden of driving,” and “transform mobility for millions of people” around the world, including those who have traditionally been unable to drive a car.

Google says its self-driving car is ready for testing on the real roads.

Google says its self-driving car is ready for testing on the real roads.

Since that announcement, Google has been hard at work refining its self-driving cars, testing them first on closed test tracks and then more recently on the private roads on the Google Campus. At the same time, Google has worked hard to refine its fleet of Lexus 450h SUV self-driving cars, being one of the first companies in the U.S. to put fully autonomous cars on the roads of California under its special autonomous vehicle licensing program.

While its prototype cars will first go on the roads with a detachable steering wheel, pedal box and fully-trained Google engineer ready and able to jump in and take control if needed, its fleet of 100 self-driving pods don’t have a steering wheel.

Google's self-driving pods will use the same on-board software as its Lexus LX450h self-driving SUVs

Google’s self-driving pods will use the same on-board software as its Lexus LX450h self-driving SUVs

On Friday last week, after success with its private road testing program for its own custom-built google cars, Chris Urmson, Director of the Google Self-Driving Car Project announced via an official blog post that Google was ready for the next phase of its development process: real-world, public road tests.

Unlike other full-size, highway-capable self-driving cars we’ve seen developed and tested by automakers like Audi, Nissan and Tesla Motors the Google self-driving prototypes are two-seat, low-speed electric vehicles with a top speed of just 25 miles per hour.  About the size of a Smart ForTwo minicar, they’ve been designed to take on duties in busy urban or suburban environments rather than inter-city freeway trips.

While that might seem like a technological easy way out, it’s not.

Unlike the autonomous driving software being developed by most automakers — which focuses on offering ‘autopilot’ functionality for high-speed freeway travel and stop/go traffic jams as well as traditional car controls for moments when a human needs to take control of the vehicle —  Google is aiming to offer 100 percent, always-active autonomy.

With so many logged miles, the driving bit of Google’s self-driving project is already pretty refined.

That means while its prototype cars will first go on the roads with a detachable steering wheel, pedal box and fully-trained Google engineer ready and able to jump in and take control if needed, its fleet of 100 self-driving pods don’t have a steering wheel.

In the future, Google’s self-driving pods will rely simply on their on-board software, and a myriad of sensors, to keep itself and its occupants safe. Despite only operating at a speed of less than 25 miles per hour, the system will need to be 100 percent infallible.

Google will first test the cars on the roads in Mountain View, California.

Google will first test the cars on the roads in Mountain View, California.

With nearly one million miles of autonomous driving logged by its fleet of autonomous LX450h SUVs thus far — which we note uses the same software as the Pod cars will use — Google says it’s amassed about 75 years of typical American adult driving experience thus far. At the time of writing, it notes that its fleet are logging an additional ten thousand miles of autonomous driving experiences per week. So far, not a single at-fault accident has been logged by any of Google’s cars, with the only accidents thus far being caused by human error, not computer error.

With so many logged miles, the driving bit of Google’s self-driving project is already pretty refined, yet Google says it’s now turning its attention to the kind of testing that can only happen on the public road. It needs to learn how the public reacts to the autonomous vehicle, as well as making some really important discoveries that we’ve never thought about — such as where a self-driving car should stop if there’s no parking at its destination or a road blockage stops it from even reaching its destination.

Like us, we’re sure you’ll want to find out.

______________________________________

Want to keep up with the latest news in evolving transport? Don’t forget to follow Transport Evolved on Twitter, like us on Facebook and G+, and subscribe to our YouTube channel.

You can also support us directly as a monthly supporting member by visiting Patreon.com.

Share on FacebookTweet about this on TwitterShare on Google+Share on LinkedInDigg thisShare on RedditEmail this to someonePin on Pinterest

Related News

  • jeffsongster

    Should be interesting to see how nerve wracking it is to ride in such a marvelous contraption. Imagine all the feet tapping imaginary brakes.

    • Joe Viocoe

      Gotta go through the birthing pains.

      We need to consult all the parents of teen aged drivers, and see how fast they coped. At least with a computer, it is consistent. Once you know how well it drives, it won’t divert into a distracted driver staring at a phone, trying to impress people, and driving drunk on weekends.

      • jeffsongster

        What scares me most is that ( as a former software test engineer) these bits of code will have us trusting them with our lives. When your word processor dies… a small part of your soul may go with it… but that is usually the biggest risk.

        When I got a car from GM ages ago and its motor control software would stall out the engine randomly in intersections and such… it was risking my life. Thankfully the car didn’t have power steering or it would have been even more dangerous a problem.

        The looming problem is that they will be so good we will trust them everyday for months… get used to only looking out the window every few minutes or so… and then one morning the click OK to continue screen we’ve grown used to will advise us about some upgrade or another… we hop in and the car drives us off a cliff because the company was desperate to ship that new feature by the next trade show to hype their stock for the next quarter… you get the idea.

        So… thrilled and terrified at the same time. Fascinated by the tech but frightened at the way our capitalist system might cut corners and drive us all off the cliff in the pursuit of faster better cheaper. Time will tell…

        • Joe Viocoe

          Yep… that is a serious concern of mine too. But I fear the drunk driver much more. The wiring in a human brain is actually much more prone to failure.

          If we truly must fling ourselves at high speeds, I want a processor with the fastest and most reliable reaction time. Failure proof doesn’t exist, but I want the best probability on my side.

        • D. Harrower

          Similar to the way we’ve seen Tesla bow to pressure from their stockholders on occassion, rather than doing what’s best for the consumer.

          • jeffsongster

            I wouldn’t single out Tesla… all public companies end up being corrupted into shareholder feeding devices rather than changing the world eventually. It is the nature of the crapitalist system. Forcing costs down, forces quality down, forces the whole mess to circle the drain eventually as we see all around the world. Complete central control is also corrupted easily. The best answers seem to be open, transparent, moderately, decentralized democratic, leadership. Keeping companies private by buying them back when low is also an excellent way of keeping the crapitalist vultures away.

  • Joe Viocoe

    Although I applaud Google’s efforts. I think the implementation is flawed. I like Tesla’s (and other automaker’s) approach. Make a normal car, and little by little integrate semi-autonomous features that get people used to it. Especially on the highway, where there are fewer variables and people want to convert their long commute into shorter segments.
    … rather than focus on small city streets, where it is harder to program a computer’s reaction to the chaos and people have to accept the whole autonomous driving thing as one ‘take it or leave it’ bundle.

  • D. Harrower

    This is certainly an exciting development that I will be watching closely.

    Even though they look a bit ridiculous, I can see these pods serving a purpose ferrying people around dense cities (though I bet there will be major blowback from cabbies when that occurs)

    The question I have is can these cars actually deliver you to your destination, or just the general area like Google Maps. ; )