South Koreans win ‘humanoid’ robot contest

Don’t try this at home: DRC-Hubo, winner of the robotics challenge, gets to grips with a valve.

A robot which looks human and can walk and drive a car has won a $2m prize in a US Defence Department competition. Is it just common sense to design robots that are similar to people?

Four years on from the Fukushima nuclear disaster in Japan, radiation levels in the area remain toxic. Around 120,000 people still cannot return to their homes and the rates of some cancers, particularly of the thyroid gland, are sharply increasing. Those tasked with clearing up the mess at the site of the disaster can face life-threatening work.

But the day when humans hand such jobs over to machines designed in our own image may be drawing nearer. At the Darpa Robotics Challenge in Pomona, California this weekend, 23 humanoid robots (robots similar to humans in appearance and actions) from around the world competed to clear an obstacle course based on Fukushima. DRC-Hubo, designed by Team Kaist of South Korea, completed all eight of the tasks set in 44 minutes and 28 seconds to secure the top prize: $2m in funding.

It was the final of three competitions run over two years by the US Department of Defence’s Advanced Research Projects Agency (Darpa). The contests had become increasingly difficult as time had progressed, with stricter time limits and the robots allowed less human help.

The robots were expected to drive alone, walk through rubble, open doors, turn valves and climb stairs. With each robot weighing around 175kg, the biggest risk was the possibility of a fall, a fate most likely when it got out of its car near the beginning of the challenge. The crowd at the event were, therefore, particularly impressed when Tartan Rescue’s Chimp robot recovered from a heavy fall without human help to finish third.

Darpa, who run high-tech projects in the spirit of discovery, say they launched the competition ‘in response to a humanitarian need that became glaringly clear during Fukushima’. They have form in stimulating interest in daring innovations: their Grand Challenges of 2004, 2005 and 2007 have been credited with spurring the development of driverless cars, prototypes of which are set to appear on Californian roads this summer.

In our own image

Jerry Pratt, an engineer with one of the teams, says it makes sense to design robots like people, especially when they’re needed in environments largely created by humans. Designing machines based on our bodies means learning from millions of years of evolution, and nature is a greater creator than we will ever be.

But Colin Angle, the CEO of iRobot, questions this, saying that walking robots are particularly impractical; robots on tracks were already available to clear up Fukushima. So maybe designing robots that look and act like us is an act of vanity. If our physical flaws require us to build machines to do our dirty work, we should aspire to make them capable of feats we could never dream of attempting.

You Decide

  1. Should we invest money in robots with similar physical characteristics to us, or without them?
  2. Is designing robots that look like us just an act of vanity?


  1. Draw a picture of a robot which you would want to send in to a disaster area, labelling its key features and explaining what it can do.
  2. Study the link to the Mashable website and create a timeline of the advances made in robotics in recent years. Choose three which you think will be particularly significant in the future and write a paragraph for each to explain why you chose it.

Some People Say...

“People are massively more versatile and adaptable than the smartest robots.”

Elaine Chen, lecturer

What do you think?

Q & A

Is it hard is it to design a humanoid robot?
Yes. The toughest element of it is programming the robot to do some of the things that most humans take for granted, like the ability to walk and to balance. It’s difficult to break these things down to mathematical formulae when we have done them largely through instinct for so long. And there are still things we do which robots can’t, such as straightening our legs.
Will they only be used in disaster areas?
They will need to be cheap and safe before being used elsewhere, but they offer a world of opportunities in the long term. We could conceivably see similar machines performing basic household tasks for us, for example. But some are also wary of robots making human jobs redundant or being used to fight wars, as pilotless planes (drones) are.

Word Watch

Fukushima nuclear disaster
On 11 March 2011, a huge earthquake (with a magnitude of 9.0 on the moment magnitude scale) happened off the north-eastern coast of Japan. It caused a tsunami with 30-feet waves which damaged several nuclear reactors and caused a meltdown at the Fukushima nuclear power plant.
Top prize
The total prize money was $3.5m, with $1m going to second-placed IHMC Robotics of Florida and $500,000 awarded to the third-placed Chimp robot, from Carnegie Mellon Unviersity in Pennsylvania.
High-tech projects
These are known as ‘moon shots’ because of their ambition, which is said to resemble that of the Apollo space programme of the 1960s, to achieve major breakthroughs on behalf of humankind.
Driverless cars
The initial Grand Challenge in 2004 went very badly: none of the 15 cars which took part succeeded. But now Google has a number of autonomous cars which collectively drive around 10,000 miles per week. Google will also launch the first self-driving cars on public roads. Several carmakers and companies such as Uber are also exploring the opportunities they offer.

PDF Download

Please click on "Print view" at the top of the page to see a print friendly version of the article.