A team of researchers at Georgia Institute of Technology says it has developed a robot that helps hospital patients slide on their gowns.
Instead of relying on robotic vision, the robot, a Willow Garage PR2, uses the forces it perceives to guide the garment onto a patient's hand, around the elbow, and onto the shoulder.
The robot taught itself to accomplish this task in a single day by analyzing about 11,000 simulated examples of a robot putting a gown onto a human arm, which allowed its neural network to learn to estimate the forces applied to the person. The robot also learned to predict the results of moving the gown in different ways, and used these predictions to select motions that comfortably dress the patient.
The team varied the PR2's timing and enabled it to think as much as one-fifth of a second into the future while considering its next maneuver, as less time than that led to more failures.
From Georgia Tech News Center
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found