PDF Basket
Therapists often use puppets or animated characters to help children with autism spectrum disorder (ASD) learn to recognise and express emotions. Although children with ASD seem to prefer interacting with non-human characters, the downside is that these props are not interactive, cannot respond to the children’s behaviour, and require intensive input from the human therapists.
Roboticists, psychologists, cognitive scientists, computer vision researchers, and ethicists involved in the EU-funded DREAM project are helping to prove that small social robots could support children with ASD.
They believe their platform will help these children engage in interactive play and improve their social skills, with the ultimate aim of achieving better real-life social interaction with other people.
Remote-controlled robots have already been designed to suit this purpose. However, the project has been working towards a smarter and more autonomous robot that can interpret children’s emotions and demands, and act accordingly to help them practice skills such as imitation, joint attention and turn-taking.
The DREAM robot platform should also be able to monitor and record the children’s behaviour. This next-generation approach – dubbed robot-enhanced therapy (RET) – will help free up therapists for other tasks.
‘Experiments assessing the usefulness of robot-enhanced therapy compared to standard therapist-assisted interventions have delivered promising preliminary results,’ says project coordinator Tom Ziemke of the University of Skövde in Sweden. ‘The project is currently carrying out the world's first large randomised clinical trial on the use of social robots in this type of psychological therapy.’
Popular and smart
The DREAM project has been adapting and modifying the humanoid, child-sized NAO (an autonomous, programmable humanoid robot), manufactured by Softbank Robotics, which is the most widely used robot in education and research in the world. It is equipped to detect and recognise pre-learned objects, faces, words and sentences, and can communicate using LED lights and a sophisticated voice synthesiser.
DREAM researchers have been programming NAO with software that will capture and analyse sensory data from the children – such as their movements, gestures, gaze, facial expressions and tone of voice. The aim is to help NAO work out how engaged children so that it can automatically adapt its behaviour to grab their attention further, when needed.
The development of ‘smart environment’ technologies is also key to the project’s success. During therapy sessions, it is envisaged that NAO will sit alongside the child and the therapist at DREAM’s custom-designed sensory table. The team has also been developing a multi-camera system that connects the robot to motion-sensing, high-resolution cameras which can track and measure the child’s movements, facial expressions and interactions with the robot.
This environment is being developed to supplement the robot’s sensory capabilities by providing it with information that would otherwise require extremely advanced perceptual and cognitive capabilities.
A rounded approach to ethics
Of course, there are concerns about safety, trust, responsibility and data use when it comes to using partially autonomous robots to interact with children. Who is responsible for the robot’s behaviour? Do parents trust the robot with their children? What is done with the data the robots collect?
The team has been taking such issues seriously by building ethical restraints and existing laws into the development and testing of their technologies.
‘The project has recently been extended by six months, so we’re currently running some additional clinical evaluations,’ says Ziemke. ‘We’re also organising two more international workshops – one on the ethical aspects of using robots for therapy.’
Engineers and designers from the technical side of the project are working closely with therapists and ethical researchers to ensure that the robot is completely suited to clinical therapy. The results are contributing to wider knowledge on robot ethics and the ethics of human-robot interaction.