Social robots are simpler to interact with than humans, can repeat games with infinite patience and record the data for further study. That’s changing the way autistic children learn new skills, say therapists.
One of the emerging applications of social robotics is in providing therapy for children with autism. Social robots are ideally suited to this task because they are simpler than humans to interact with, their actions are perfectly repeatable and they can be modified in various ways to meet the requirements of different children.
So therapists have begun to test these devices when working with autistic children. And today, John-John Cabibihan at the National University of Singapore and a couple of pals review this work, the numerous different types of robots that therapists are using and the techniques they have developed to engage and influence autistic children. The result is a useful introduction to an area of social robotics that has significant promise.
Autism is a developmental disorder that encompasses a wide spectrum of impairments in social skills, communication and imagination. It is a life-long disorder but there is growing evidence that early intervention can make a significant difference to children’s later quality of life.
Social robots are well suited to this kind of intervention. To start with, they are useful for diagnosing autism, which usually, autism cannot be diagnosed until a child is at least three years old. But there is growing evidence that a diagnosis can be made much earlier by studying eye contact in high risk babies (those with close relatives who are autistic, for instance).
Social robots, with cameras for eyes, can monitor eye contact accurately and over long periods of time to gather evidence for a diagnosis one way or the other.
Another application is in eliciting certain types of behaviour. Social robots can play a central role in exercises that help children improve their social, sensory and cognitive skills as well as their motor control. Therapists have programmed the robots to perform tasks that help the child practice eye contact, turn taking, imitation and so on. “Such activities include teaching a child to initiate greetings, to wait for its turn to throw the ball, to follow the robots gaze to an object of interest, and to copy the robots movements as it dances,” say Cabibihan and co.
In the process, the robot becomes a friendly playmate and a mediator between the therapist and the child, difficult roles for humans to play consistently.
Cabibihan and co also review the wide range of robots that therapists have pressed into service. These range from hyper-realistic humanoids such as FACE (Facial Automation for Conveying Emotions) to entirely non-human robots such as Roball. It’s an impressive list.
The clear message from Cabibihan and co is that social robotics is revolutionising the way experts diagnose, study and help autistic children.
There are certainly challenges ahead of course. For example, these guys point to the need to better characterise the change in a child’s behaviour towards real people as a result of social robotic therapy. “This is very important since the very purpose of therapy is to facilitate the child’s social interaction with other people, not just with the robots,” say Cabibihan and co.
And there is clearly much to be learned in how best to use social robots for children with different needs.
Nevertheless, the data gathered from this type of work has the potential to change the way we think about autism and, most important of all, to improve the outcomes for children with autism.