Then and "Nao": UD lab experimenting with social robot's use in child rehabilitation
University of Delaware researchers are exploring a social robot’s application to pediatric rehabilitation in collaboration with Johns Hopkins University.
At UD’s Spencer Lab, foam flooring, steps and boxes are set up to create a playground where children with motor disabilities can interact with Nao, a social robot.
UD’s Herbert Tanner is one of the researchers involved in the project, which he said is about engaging children ages one to five in social learning.
"There are some kids that we have observed in pilot studies that feel very comfortable around Nao," Tanner said. "Other infants, little kids, maybe are sort of a little intimidated because they’re looking at a moving machine that’s about their own size. So we’re trying to throw into the mix different types of robots, more children friendly, more colorful perhaps; we want to make the whole environment as child-friendly as possible."
The team, which also includes UD's Cole Galloway and Jeffrey Heinz, recently received a grant from the National Institute of Health to explore how Nao, along with a harness system needed to support these children, can be used in pediatric rehabilitation.
Tanner said they’re focusing on ages one to five in order to start rehabilitation as early as possible. Once a child hits age five, weight restrictions with the harness system used to support these children come into play.
The harness system supports a child’s weight and helps them move freely. It consists of a metal frame designed with pulleys that make it easier for a child to stand up.
Nao can stand up, sit down, raise its arms, and talk to people, among other functions. Tanner said it can be programmed to use sounds and verbal signals to interact and assist children with motor skills and response. Nao can also walk at a slow pace towards a target — like a game of chase.
Prasanna Kannappan, a PhD student studying mechanical engineering, said there are multiple sensors in the lab that will be used to capture the motion of the child — that researchers will be able to analyze later.
“This particular project — children are very unpredictable and it’s interesting to try to model or predict what they’re going to do,” Kannappan said. “That will be interesting in that sense. There are a lot of computer vision problems where we try to identify what others are doing, and it’s a lot more challenging to solve the problem. It’s interesting to look at things like that.”
Tanner added the team is trying to come up with additional high-level behaviors for Nao.
"We need to direct its motion, tell it where to go and how fast — what sequence of maneuvers or motions would be meaningful in interactions with a child," Tanner said.
Nao can be programmed to use sounds and verbal signals to interact and assist children with motor skills and response.
And Nao is just one piece of a larger puzzle, Tanner said.
"Nao might be in the spotlight at this moment in time but there are so many things that go around Nao," Tanner said. "Nao, for us, is just a tool and it’s a component in a bigger system that involves cameras that capture what the child is doing, an activity recognition system that helps us understand, almost in real time, what’s happening in the scene."
Tanner said the ultimate goal is develop less expensive alternatives to Nao for real world use in classrooms, clinics and home environments.