October 31, 2012      

To take on and succeed at DARPA’s latest Grand Challenge, a robot better have a the lower torso mobility of an NFL running back or it may as well stay home.

Here’s what that robot and its nimble torso needs to accomplish:

“Drive a car; walk over an uneven, debris-strewn surface; climb shaky industrial ladders and catwalks; use power tools to break through a concrete panel; find and close a valve near a leaky pipe; and replace a piece of industrial machinery like a cooling pump.”

The arms, hands and shoulders needed for the above tasks are much further along than robotic developments for the lower torso. Humanoid robots contesting for the DARPA prize must also be fleet afoot.

Giving robots the tendons and musculature of Homo sapiens, “has been a no-no for some time” among robot engineers, said Dennis Hong, associate professor and the founding director of RoMeLa (Robotics & Mechanisms Laboratory) of the Mechanical Engineering Department at Virginia Tech.

Dennis Hong and ASH

Hong has his own mobile, lower-torso favorite, ASH, Autonomous Shipboard Humanoid, the latest from his RoMeLa team being built for the Office of Naval Research. ASH is based on Hong’s award-winning CHARLI-2 mobile robot. However, ASH is going to have a major upgrade from CHARLI-2, mainly, titanium springs in his legs and a butt that performs like human muscles.

This Grand Challenge is the third for DARPA in recent years.

“The first three focused on self-driving cars, were wildly successful in acceleration the development of autonomous vehicles,” reported Wired. “Not long ago, the idea of self-driving cars that could share the road with humans seemed like fantasy.”

Not any longer, as, for example, the Google autonomous car zooming over Nevada’s highways recently.

Industry observers hope that a little attention from DARPA will do for humanoid robotics what it did for robot vehicles.

Hong’s ASH is in the running, so too is Boston Dynamics (the folks who brought us the amazingly agile PETMAN and BigDog).

Boston Dynamics’ PetProto, a relative of PETMAN, seems to have the skills DARPA is looking for. However, even PetProto won’t be Boston Dynamics’ ultimate entry. It’s raising the bar on agility yet again with its as-yet unfinished Atlas machine.

Atlas is a very close but more advanced relative of PetProto that Boston Dynamics will use as a test bed to further polish the skills necessary to challenge for the DARPA crown.

Other teams throwing their hats into the ring: RE2, the University of Kansas, Carnegie Mellon, the Massachusetts Institute of Technology, TRAC Labs, the University of Washington, the Florida Institute for Human and Machine Cognition, Ben-Gurion University, NASA’s Jet Propulsion Laboratory and TORC Robotics.

See related:
Four U.S. Agencies Announce $40M In Research Grants For Next-Gen Robotics
DARPA to Fund Multiple $5M Projects to Improve Energy Efficiency in Robots