acm-header
Sign In

Communications of the ACM

ACM TechNews

Robots Can Now Teach Each Other New Tricks


View as: Print Mobile App Share:
A robot teacher.

Two robots have demonstrated the basic ability to teach each other.

Credit: Shutterstock.com

Two robots operating in different academic research labs have demonstrated a rudimentary ability to exchange knowledge.

The PR2 research robot at Cornell University was first trained to perform a task via the RoboBrain database, and the Baxter robot at Brown University--a different type of machine--took PR2's learned knowledge and used it to determine how to perform the same operation in its own environment.

Information-sharing by robots could reduce the need for significant reprogramming, as well as enabling robots to adapt quickly when confronted with a new task or an unfamiliar setting. "When you put a robot in a new situation--and in the real world it happens in every room the robot goes into--you somehow want that same robot to engage in autonomous behaviors," says Brown professor Stefanie Tellex.

The physical differences between the two robots meant low-level commands would not match, so Tellex's team needed to work out a scheme to enable command transfers between the two platforms.

Tellex says the goal is to enable a robot to determine how to translate data for itself, based on how its physical configuration compares with that of another robot.

Cornell professor Ashutosh Saxena expects robots to increasingly share knowledge in the future.

From Technology Review
View Full Article

 

Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account