DARPA's Atlas

On Thursday, DARPA unveiled one of the more frightening and futuristic robots we've ever seen. But the massive 6-foot humanoid robot, which was developed by Boston Dynamics, isn't designed to kill -- it's here to help us.

Atlas, with 28 hydraulically actuated joints, is one of the most advanced humanoid robots ever built. But the humanoid is essentially a physical shell, a starting point for the software brains and nerves that the teams from DARPA’s Virtual Robotics Challenge will use as a development platform.

We're not yet in danger from these bipedal bots, but better sensors, muscle-like actuators, and more advanced planning and control systems are making robots more and more like us everyday.

After more than 30 years of robotics development, we are soon sure to see the machines walking among us.

Take a look at the evolution of robotics development.
Updated:
Photo by: DARPA/Boston Dynamics / Caption by:

Rubber artificial muscle, 1967

In 1967, an artificial muscle was developed by professor Ichiro Kato and his team at Waseda University in Japan, one of the centers of early robotics research. Rubber embedded in several threads along a longitudinal axis for the first time created muscle-like movements which contracted longitudinally due to lateral constraint by the threads.
Updated:
Photo by: Humanoid Robotics Institute, Waseda University / Caption by:

Master-slave walking machine, 1968

Using a mechanical model of lower-limbs movement, Waseda University's WL-3, which had an electro-hydraulic servo-actuator and was controlled by using a master-slave method, was constructed.

The WL-3 for the first time managed human-like movement in a swing and a stance phase which also allowed the limbs to stand up and sit down.
Updated:
Photo by: Humanoid Robotics Institute, Waseda University / Caption by:

Artificial muscle made of rubber, 1969

The anthropomorphic pneumatically activated WAP-1 had artificial muscles made of rubber which were attached as actuators. Bipedal movement was made possible by teaching playback control of its artificial muscles.
Updated:
Photo by: Humanoid Robotics Institute, Waseda University / Caption by:

Static Walking, WL-5, 1970

Controlled with a minicomputer, Waseda University's WL-5 could walk and change direction, though it moved at the slow speed of 45 seconds per step.
Updated:
Photo by: Humanoid Robotics Institute, Waseda University / Caption by:

Wabot-1, 1973

Wabot-1 was the first full-scale anthropomorphic robot built in the world, integrating a limb control system, a vision system, and a communications platform.

Sensors allowed Wabot-1 to measure distances and directions to objects, and the robot could walk and even grip and move objects with hands that used tactile sensors.
Updated:
Photo by: Humanoid Robotics Institute, Waseda University / Caption by:

Realization of Dynamic Walking, 1984

Waseda University's WL-10RD integrated an ankle and hip joint which provided feedback that enabled the robot to better transition its balance from one leg to the other.

This advancement in technology made walking possible for the first time, albeit at just 1.3 seconds per step.
Updated:
Photo by: Humanoid Robotics Institute, Waseda University / Caption by:

Shadow, 1987

Shadow, founded by Richard Greenhill in 1987, has been making robots and developing associated technologies that aim to build a general-purpose robot that might be useful in homes to perform daily tasks.
Updated:
Photo by: Shadow / Caption by:

Manny, 1989

Pacific Northwest National Laboratory robot Manny was built for the U.S. Army in 1989. Manny was life-size and very human looking for the time, but had no intelligence or autonomous mobility.
Updated:
Photo by: Pacific Northwest National Laboratory / Caption by:

P2, 1996

Honda's battery-powered 6-foot, 460-pound P2 android was introduced in Tokyo on December 20, 1996.
Updated:
Photo by: Honda / Caption by:

H5, 1998

Jouhou System Kougak laboratory at Tokyo University built the full-size H5 android in 1998.
Updated:
Photo by: Tokyo University / Caption by:

REEM-A, 2005

REEM-A, a walking, manipulation, speech and vision development platform, was completed in 2005. In 2006 it was ready to take part in the RoboCup competition in Bremen where it won the walking competition and was a semifinalist in the penalty kicks portion of the games.

In 2007, using the same software as 2006, REEM-A also took part in the RoboCup in Atlanta and was once again a finalist in the penalty kick competition.
Updated:
Photo by: PAL Robotics / Caption by:

Robothespian, 2005

RoboThespian, which began in England in 2005 from Engineered Arts Limited, is fully interactive, multilingual, user-friendly, capable of communicating and interacting with humans, and available now for around $85,000.
Updated:
Photo by: Engineered Arts Limited / Caption by:

HRP-3 Promet, 2007

HRP-3 Promet Mk-II, HRP-2 Promet, and HRP-3 Prototype humanoid robots are displayed during the Promet Mk-II press preview at Kawada industry's laboratory in Haga town near Utsunomiya city north of Tokyo, June 21, 2007.

The 160cm-tall and 68kg humanoid robot is the upgrade model from HRP-2 Promet humanoid robot jointly developed by Kawada Industry, National Institute of Advanced Industrial Science and Technology, and Kawasaki heavy industry.
Updated:
Photo by: TOSHIFUMI KITAMURA/AFP/Getty Images / Caption by:

Dexter, 2007

Anybots in Mountain View, Calif., has developed two humanoid robots, Dexter and Monty.

Dexter, seen here jumping, has two legs, stands 5-foot and 10-inches tall, and weighs 135 pounds. He can walk, jump, and dynamically balance his robot body to remain standing just like a person does.

Anybots created a system in the robot's body that distinguishes up from down. The system uses that information to constantly adjust the "muscles," or air cylinders, in the robot's legs.
Updated:
Photo by: James Martin/CNET / Caption by:

Aero-Blue, 2008

Humanoid robot Aero-Blue picks up a T-shirt from a laundry basket to fold it during the housekeeping robot contest in Kawasaki in Kanagawa prefecture, suburban Tokyo on August 17, 2008. Robots are increasingly capable of performing day-to-day tasks.
Updated:
Photo by: YOSHIKAZU TSUNO/AFP/Getty Images / Caption by:

Honda's Asimo

Asimo, Honda's humanoid robot, appears during the FIRST Championships at the America's Center on April 28, 2011 in St. Louis.
Updated:
Photo by: Dilip Vishwanat/Getty Images for Honda / Caption by:

Physical Intelligence

With better control systems, sensors, and life-like actuators, these robots are already capable of navigating advanced terrain and using tools designed for humans.

What's next? Why, brains, of course -- and DARPA is working on that, too.

Earlier this year in National Defense Magazine, James K. Gimzewski, professor of chemistry at the University of California at Los Angeles, said that by mimicking self-organizing traits of the brain, we may soon be looking at a revolutionary breakthrough in robotic systems.

Gimsewski has been working with DARPA on a program called Physical Intelligence.

“Rather than move information from memory to processor, like conventional computers, this device processes information in a totally new way," he said.

Updated:
Photo by: DARPA/Boston Dynamics / Caption by:
Hot Galleries

CNET's Holiday Gift Guide

'Tis the season for a gadget upgrade

Check out these 8 tablets you'll want to bring home for the holidays.

Hot Products