| Meet Speedfolding The Shirt Folding Robot |
| Written by Kay Ewbank | |||
| Sunday, 06 November 2022 | |||
|
An international team of robotics researchers has developed a new robot for folding laundry. SpeedFolding is described as being reliable and efficient. Whether the 'speed' bit of the name is justified probably depends on what you're measuring it against. What makes the new robot different to previous attempts to develop a clothes folding robot is the design based on two arms. The researchers, from UC Berkeley and the Karlsruhe Institute of Technology, will present a paper on the robot and the neural network that underpins it at a robotics conference in Kyoto this month.
The team says that folding garments reliably and efficiently is a long standing challenge in robotic manipulation due to the complex dynamics and high dimensional configuration space of garments. SpeedFolding is given user-defined instructions as folding lines, and based on these takes a garment, smooths and folds it. The system is based on a novel neural network architecture that is able to predict pairs of gripper poses.
The network was trained using 4300 human-annotated and self-supervised actions, and is able to fold garments from a random initial configuration in under 120 seconds on average with a success rate of 93%, and can fold around 30 to 40 garments an hour. While this is slower than a well-trained human would be able to fold laundry, it is much faster than previous attempts using robots, and the research team plans to explore methods that can learn to manipulate a novel garment given a few demonstrations.
More InformationConference Paper On SpeedFolding
Related ArticlesRobot Dog From Rolling On Floor To Walking In One Hour A Robot Finally Learns To Walk Meet Ameca - The Future Face of Robotics Pieter Abbeel Wins 2021 ACM Prize In Computing A Robot Learns To Do Things Using A Deep Neural Network
To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin. 
Comments
or email your comment to: comments@i-programmer.info |
