The Georgia Institute of Technology has addressed the growing need for individuals who need assistance getting dressed because of injury, sickness or age. Their robotic-armed machine helps slide hospital gowns on people’s arms. It relies on the weight and force it feels as it guides the gown onto the arm, around the elbow and through the hand, according to Science Daily.
The robot, a PR2, analyzed over 11,000 simulated examples of putting a gown onto an arm. These examples included complete failures and total successes. Sometimes, if the gown was stuck on the arm the robot exerted too much force to get the gown over the arm. Through these practice rounds, the PR2 neural network learned what forces it needed to use in order to successfully get the gown over a person with ease.
“People learn new skills using trial and error. We gave the PR2 the same opportunity,” said Zackory Erickson, lead Georgia Tech Ph.D. student on the research team, according to Science Daily. “Doing thousands of trials on a human would have been dangerous, let alone impossibly tedious. But in just one day, using simulations, the robot learned what a person may physically feel while getting dressed.”
After the simulations, the robot practiced dressing actual people. The robot was able to use its sense of touch to strategize what force and movements would have the best results for dressing the patient.
“The key is that the robot is always thinking ahead,” said Charlie Kemp, associate professor in the Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and Emory University, according to Science Daily. “It asks itself, ‘if I pull the gown this way, will it cause more or less force on the person’s arm? What would happen if I go that way instead?'”
Currently, the robot can successfully put the gown on a person’s arm in about 10 seconds. In the future, the team hopes this development turns into successfully dressing a full person.