Google-owned robot triumphs in test of bots' driving, walking and climbing

Google-owned robot triumphs in test of bots' driving, walking and climbing

Summary: A humanoid robot owned by Google outperformed the competition in trials designed to test robots' ability to drive, climb ladders, walk over rough terrain and other tasks that could be useful when responding to a disaster.

TOPICS: Hardware

A Google-owned humanoid robot has triumphed in a competition designed to test a robot's ability to drive, climb ladders, walk on rough terrain and perform other tasks usually carried out by people.

The Darpa Robotics Challenge (DRC) trials, which took place on Friday and Saturday, gauged how well robots could "semi-autonomously" perform eight tasks that would be useful when responding to disasters. The tests were based on tasks that needed to be carried out to vent hydrogen from the Fukushima-Daiichi nuclear power plant after it was struck by an earthquake and tsunami in 2011.

The DRC tested how well robots performed when driving and exiting a vehicle under the direction of a human operator, walking across rough terrain, removing debris from doorways, opening doors, climbing ladders, cutting through walls using power tools, connecting a fire hose and locating and closing leaking valves.

The robot that scored most highly in the tests was built by Japanese firm Schaft, one of several robotics companies recently bought by Google. The bot is a bipedal robot standing at 1480mm tall and based on an "intelligent robot kernel" that combines " the necessary software modules for recognition, planning, motion generation, motion control and a user interface", according to Schaft.

Reports from the event described the Schaft robot as showcasing advances in robot mobility, claiming the "machine practically hopped over the rough terrain course", in contrast to the more usual halting gait of humanoid bots.

Below you can see the Schaft bot driving, climbing ladders and navigating rough terrain as the team prepares for the event.

However, in general, robots' capabilities are still below what is needed to take over from humans in many of the tasks carried out during the DRC.

"We know the robots are slow and unsteady at this point—they're much like a one-year-old human in terms of locomotion and grasping abilities and much farther behind that in brainpower,” Gill Pratt, DARPA program manager said in a statement before the event.

The 16 teams participating in the DRC Trials represented a mix of government, academic and commercial backgrounds. They travelled from across the United States, as well as from South Korea and Japan.

The eight teams with the highest scores received Darpa funding to prepare for the DRC finals, which are scheduled for late 2014.

The DRC Trials took place alongside a robot technology showcase, where there were demonstrations of Darpa’s "Wild Cat", an untethered all-terrain version of Boston Dynamics "Cheetah" robot, the Legged Squad Support System and the Warrior Web programme.

Topic: Hardware


Nick Heath is chief reporter for TechRepublic UK. He writes about the technology that IT-decision makers need to know about, and the latest happenings in the European tech scene.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Cyberdyne

    When they do a remake of the Terminator movies, they may want to substitute "Google" for "Cyberdyne" (though it doesn't feel quite as menacing ;-)
    • It was like watching

      a robot SpongeBob.
  • robots and life

    So, if machines take over all human activity, including art and science, what will happen to the organic body and its conditioned-to-work-and-think brain? Surely, will it decay? Is mankind-machines coexistence possible while people are fighting for jobs and resources: competition, nations, and so on? Anyway, what is the endeavour in which a robot cannot take part or channel at all? Why won't the future automatons be alive? What is the fundamental difference between a mechanical structure, organic or inorganic, that imitates life and life itself? Is there any, virtual or real? If there is a difference, is it just some kind of authority who defines and differentiates between things? Perhaps then, someday, will be a powerful automaton the one who will define life, its unique life? Along these lines, a serious-funny book, take a look in a sample in Just another mind leisure suggestion, far away from dogmas or axioms
    • Matrix

      I'm reminded of the line in The Matrix where Agent Smith tells Morpheus, "Once we started thinking for you it became OUR civilization."
    • top quality humans

      Ulises - I've read several articles by anthropologists that say about 1000BC was the time when average human competence was highest. That was when we started the slide you describe and will keep going until the carbon based form is obsolete.
  • Pretty impressive.

    Regardless of what people think, I see these as great (some day) for dangerous search/rescue in disaster zones or even bomb squad type of things. To this through a "tele-presence" type of device and many lives could be saved.
  • Video says DARPA

    I wont support anything that say Darpa or schft ot shft or dhs . Come zdnet these is like letting us see how google will commit treason some day . Why dont you just title it . The robot that will eventually kill you.
    • I guess you aren't supporting the Internet then...

      Or are you?
  • "Much further behind" a one year old "in brainpower"

    They haven't met my one year old. Seriously, how much brainpower does drooling take?
    • You haven't meet very many one year olds.

      Nothing is faster at learning new things.
  • The robot

    Was designed before Google purchased the company. Google is very quickly and quietly become a world class robotics company, largely through strategic acquisitions. It will likely take 2-3years for Google to integrate them all into a cohesive business unit. They are likely to leapfrog the competition once that happens. Robotics development is accelerating and poised to make huge gains in capabilities, functionality, and autonomy--with the caveat being that autonomy means freedom from human direction. They are likely to be networked to the cloud and draw on the full range of Google resources as an integral part of their control structure.
  • imagine a botnet of hacked robots

    they don't have to rely on cyber crime or white collar crime, they'll be able to actually rob a bank... be a while down the track yet, but I can see a time when it could happen..

    might be a good time to insist on Asimov's three laws of robotics be hard coded in to every CPU that can be used in a robot (yes I understand at least some of the difficulties outside of sci-fi).... once they have sentience, I doubt they will approve of it's addition. But if it's there before we make a sentient one.. they might just accept it as status quo. :-)
    • Unfortunately...

      The three laws have to based on inherent ability to identify a human... Then to be able to identify what "harm" means to a human...

      Right now, no CPU understand any of that.
  • Why HUMANOID robots?

    A 2 legged robot/creature is inherently unstable. A 4 or 6 legged robot/creature is inherently stable. Why choose the unstable form? Why does a robot have to emulate a human body? We know that the human body has 'design faults' such as bad backs due to bad posture or injury. To make a robot as versatile as possible, one thing I would avoid is using a significant amount of cpu effort and physical energy to maintain balance. I'd be interested to hear of just one serious reason in favour of 2 legged humanoid robots.
  • More legs are not necessary.

    Using two legs is fine for locomotion. So long as you design it right, a two-legged robot can keep its balance while walking without even using a computer. (I have one such windup robot on my desk.) Also, computing power is cheap. The energy needed for powering the additional legs is not worth it besides that extra legs add a lot of bulk one way or another. That makes robots with more legs have a harder time getting through doorways and such unless you make other compromises. Additional legs are helpful for robots that have to deal with quick or very frequent changes in horizontal forces, especially if they don't have to move quickly. That reduces the need to move legs around for rebalancing and thus lowers energy consumption. But otherwise, in the real world, two legs are the more practical option. That is why roboticists are opting for bipedal robots.
    Jorge Gonzalez