empty home about sciencentral my account help contact shim
sciencentral news
shim shim shim physical sciences shim technology shim full archive
shim shim biology genetics health brain learning animals psychology
December 22, 2004
shim
shim
shim
ScienCentral, Inc. home
  Robo Dancer    
Search
 

Smart Robots (04.24.03) - Scientists are making robots that can smile, frown, and even react to your expressions and emotions.

Mind Control (9/24/03) - Brain researchers are turning thoughts into commands.

 

Robocup

Kismet the Robot



   03.23.04
email to a friend
 
 
play video Video
image: Artspectrum
(movie will open in a separate window)
Choose your format:
Quicktime
Realmedia

If ballroom dancing or swing is your thing, here's a new twist on how dance partners communicate. This ScienCentral News video has more.

Bust a Move

Want to dance, but don't have a partner? You can always try programming one.

Sommer Gentry, robotics scientist at the Massachusetts Institute of Technology and competitive swing dancer, used her dancing expertise to test haptic signaling—or the sense of touch— between humans and robots. Dancing that involves a leader and a follower relies on the coordination of actions between the haptic signals that the partners give each other.

"There are two things really that go into dancing with another person, and one of them is knowing the vocabulary of movement," explains Gentry. "All of my friends that I swing dance with, we know the same basic movements. There's a haptic signal, a slight push or pull on your hand or your back that tells you which of these moves that I know I am going to do now."

Gentry programmed an armlike robotic device, called the PHANToM, to move to the strains of "New York, New York." Then she designed a human-robot dance that was made up of four moves—clockwise circles, counter-clockwise circles and half circles in each direction. "The idea was to have a human-robot interaction that was similar to swing dancing in that it was composed of movements that the person was familiar with in an unknown order," explains Gentry. "The person knows that there are these four possible moves and then the robot signals can be interpreted as only one of those four moves. What I wanted to show was that people, when they were dancing with my robot, were using the fact that they know the move."

The PHANToM
image: Sommer Gentry and Dorry Segev
Gentry says it normally takes a tenth of a second for the brain to process a haptic signal such as a dance lead. But when following a robotic leader, test subjects started their next move twice as fast—within a twentieth of a second. "There's no way that they could have felt the robot tell them to turn around before they actually did turn around," says Gentry. "They must have been preplanning that turnaround, which shows that they must have been using the knowledge of the movement. They knew that the only move that's like that has a turnaround coming up. They're using this move vocabulary to interact with the robot better to get lower delays than would be possible if we didn't have a move vocabulary."

Gentry believes her study could some day help robotic surgeons, like her husband and regular dance partner Dorry Segev. Chief resident of general surgery at Johns Hopkins Hospital, Segev says that right now in robotic surgery, "there's no touch feedback, there's no haptic feedback from the robot. It's very easy in the early parts of the training with the robotic instrument to tear tissue, because the only feedback you get from the robot is visual, and you can't feel when a tissue is giving back resistance. With more advanced robot-human interaction, the robot could actually help you by wielding another instrument sort of in harmony with what you're doing inside the patient as a first or second assistant. In addition, the robot could more than just give you the real haptic feedback that's inside the patient, it could maybe give you more intelligent haptic feedback. So, if the robot can sense the presence of a structure that you would ideally want to avoid injuring, then the robot would provide more directed feedback to make it less likely that your instruments would come near that structure."

This research was presented at the 2003 IEEE International Conference on Systems, Man, Cybernetics, and was funded by the Department of Energy Computational Science Graduate Fellowship.


 
top       email to a friend by Karen Lurie


Terms of Use     Privacy Policy     Site Map     Help     Contact     About     My Account
 
ScienCentral News is a production of ScienCentral, Inc. in collaboration with the Center for Science and the Media 248 West 35th St., 17th Fl., NY, NY 10001 USA (212) 244-9577. The contents of these WWW sites © ScienCentral, 2000- 2004. All rights reserved. The views expressed in this website are not necessarily those of the NSF. NOVA News Minutes and NOVA are registered trademarks of WGBH Educational Foundation and are being used under license. Image Credits National Science Foundation
ScienCentral home