text-only page produced automatically by LIFT Text Transcoder Skip all navigation and go to page contentSkip top navigation and go to directorate navigationSkip top navigation and go to page navigation
National Science Foundation
News
design element
News
News From the Field
For the News Media
Special Reports
Research Overviews
NSF-Wide Investments
Speeches & Lectures
NSF Current Newsletter
Multimedia Gallery
Search Multimedia
Image
Video
Audio
More
Multimedia in the News
NSF Executive Staff
News Archive
 

Email this pagePrint this page
"Robo-Babies" -- The Discovery Files


The Discovery Files
Audio Play Audio
The Discovery Files podcast is available through iTunes or you can add the RSS feed to your podcast receiver. You can also access the series via AudioNow® by calling 405-875-0058 on any telephone.

In a first-of-its-kind experiment, a roboticist from the University of Vermont created both simulated and actual robots that change their body forms while learning how to walk. And, over generations, the roboticist's simulated robots also evolved, spending less time in "infant" forms and more time in "adult" four-legged forms.

Credit: NSF/Karson Productions

Audio Transcript:

Baby 'Bots (Sound effect: robotic baby cry) They Just Grow Up so Quickly.

I'm Bob Karson with the discovery files -- new advances in science and engineering from the National Science Foundation.

Most of the time, when engineers set out to build a robot, they design fixed body parts and program the robot's behavior. But new experiments at the University of Vermont show there may be a better way: build a robot baby and let it develop its behavior -- learning and adapting while its body is "growing up" and changing shape.

(Sound effect: toy robotics sounds) Roboticist John Vongaurd conducted experiments using simple Lego mindstorm kit robots in over 5,000 computer simulations. The challenge for the robots was to make their way over to a light source without tipping over. The creatures experimented with various motions, depending on their body form at the time -- slithering, shuffling and eventually walking. To Vongaurds' surprise, these 'evolving' robots learned to walk sooner than the ones with fixed body parts, and were better at challenges they hadn't faced before -- like being knocked with a stick.

In simulated models, the robots demonstrated 'learning' patterns similar to a human infant (Sound effect: happy baby sound), moving, then balancing, then refining motions based on interactions between its brain and its body.

The experiments demonstrated for the first time that learning to walk while flat, squat and then upright, makes robots better walkers than those designed to walk upright from the start.

I'm hoping if I leave enough clothes on the floor, my robotic vacuum cleaner will eventually learn to do laundry. Like that's gonna happen.

"The Discovery Files" covers projects funded by the government's National Science Foundation. Federally sponsored research -- brought to you, by you! Learn more at nsf.gov or on our podcast.

 
General Restrictions:
Images and other media in the National Science Foundation Multimedia Gallery are available for use in print and electronic material by NSF employees, members of the media, university staff, teachers and the general public. All media in the gallery are intended for personal, educational and nonprofit/non-commercial use only.

Images credited to the National Science Foundation, a federal agency, are in the public domain. The images were created by employees of the United States Government as part of their official duties or prepared by contractors as "works for hire" for NSF. You may freely use NSF-credited images and, at your discretion, credit NSF with a "Courtesy: National Science Foundation" notation. Additional information about general usage can be found in Conditions.

MP3 icon
NSF podcasts are in mp3 format for easy download to desktop and laptops, as well as mobile devices capable of playing them.

 



Email this pagePrint this page
Back to Top of page