Tech
Knife-Wielding Robot Learns How to Handle Groceries with New Technique (Video)
Catherine Griffin
First Posted: Nov 05, 2013 09:02 AM EST
Give a robot a knife, and it won't know what to do with the tool. Yet researchers from Cornell University are teaching one robot exactly how to work with a blade. They've worked with a Baxter robot from Rethink Robotics in Boston and have engineered it to "coactively learn" from humans and make adjusts while an action is in progress.
Modern robots, like those on automobile assembly lines, have no brains--just memory. An operator programs the robot to move through the desired action and the robot repeats the exact same action every time an object goes by. This works well when a robot doesn't need to adapt and is on an assembly line, but things become slightly more complicated when a machine is taken away from this assembly line.
Researchers are looking into creating personal robots for the home. This, in particular, means that repetitive actions won't cut it. For example, a robot has to learn to handle tomatoes more gently than canned goods. If it picks up a kitchen knife, it should be smart enough to keep the blade away from humans.
In this case, the Baxter robot has the ability to plan its own movements. It displays three possible trajectories on a touch screen where the operator can select the one that looks best. In addition, humans can give corrective feedback. As the machine executes its movements, a person can intervene and manually guide the arms to fine-tune the trajectory. The robot can then learn from this and perform better the next time.
"We give the robot a lot of flexibility in learning," said Ashutosh Saxena, one of the researchers, in a news release. "The robot can learn from corrective human feedback in order to plan its actions that are suitable to the environment and the objects present."
In fact, the robot can begin to associate a particular trajectory with each type of object. For example, a quick flip over might be the fastest way to move a cereal box, but it wouldn't work with a carton of eggs. The robot is taught to handle eggs more carefully since they're fragile.
The findings are published online here.
Want to see the robot move for yourself? Check out the video below, courtesy of YouTube.
See Now:
NASA's Juno Spacecraft's Rendezvous With Jupiter's Mammoth Cyclone
©2024 ScienceWorldReport.com All rights reserved. Do not reproduce without permission. The window to the world of science news.
More on SCIENCEwr
First Posted: Nov 05, 2013 09:02 AM EST
Give a robot a knife, and it won't know what to do with the tool. Yet researchers from Cornell University are teaching one robot exactly how to work with a blade. They've worked with a Baxter robot from Rethink Robotics in Boston and have engineered it to "coactively learn" from humans and make adjusts while an action is in progress.
Modern robots, like those on automobile assembly lines, have no brains--just memory. An operator programs the robot to move through the desired action and the robot repeats the exact same action every time an object goes by. This works well when a robot doesn't need to adapt and is on an assembly line, but things become slightly more complicated when a machine is taken away from this assembly line.
Researchers are looking into creating personal robots for the home. This, in particular, means that repetitive actions won't cut it. For example, a robot has to learn to handle tomatoes more gently than canned goods. If it picks up a kitchen knife, it should be smart enough to keep the blade away from humans.
In this case, the Baxter robot has the ability to plan its own movements. It displays three possible trajectories on a touch screen where the operator can select the one that looks best. In addition, humans can give corrective feedback. As the machine executes its movements, a person can intervene and manually guide the arms to fine-tune the trajectory. The robot can then learn from this and perform better the next time.
"We give the robot a lot of flexibility in learning," said Ashutosh Saxena, one of the researchers, in a news release. "The robot can learn from corrective human feedback in order to plan its actions that are suitable to the environment and the objects present."
In fact, the robot can begin to associate a particular trajectory with each type of object. For example, a quick flip over might be the fastest way to move a cereal box, but it wouldn't work with a carton of eggs. The robot is taught to handle eggs more carefully since they're fragile.
The findings are published online here.
Want to see the robot move for yourself? Check out the video below, courtesy of YouTube.
See Now: NASA's Juno Spacecraft's Rendezvous With Jupiter's Mammoth Cyclone