We are happy to host Dr Giuseppe Palestra this month for an interview at ASDTech. Dr. Palestra is a researcher in Social Robotics and Assistive Techologies currently working at the Institute for Intelligent Systems and Robotics in Sorbonne University. Ηe spoke to ASDTech about clinical applications of social robots in autism treatment.
GP(Giuseppe Palestra) The humanoid social robots most commonly used with autistic children are NAO and MILO. Both are approximately 50cm in height, short enough to not scare the child and easy to interact with even for children with restricted mobility. Those robots offer opportunities for interactions simpler than with an actual human being in a therapeutic environment. Robots in a way become mediators in the human-human interaction.
GP: Yes, they usually respond very well. I have facilitated a great deal of experiments working with more than 90 children and I can only recall two or three instances when a child hesitated to approach the robot. Autistic children are usually curious about the robots and their capacities as they often have a strong affinity for electronics and technology in general.
GP: We still need to improve the system’s recognition of autistic children’s speech, as that can be quite idiosyncratic. More work is also required in the actual social interaction between the child and the robot, more precisely the recognition of the behavioural patterns of the children.
GP: The role of the therapist during the child-robot interaction is of fundamental importance. The robot is not there to replace the therapist but it is rather a mediator, a tool selected and approved by the therapists that treat the child.
GP: We try to use metrics as objective as possible. For autistic children, we try to quantify behaviours that pertain to their deficiencies. For example, during a therapy session, we log how often the child makes eye contact, the number of positive and negative facial expressions, the direction of the gaze, the position of the child in relation to the robot and the room, the body posture, the number of correct and incorrect answers to the robot’s questions and the time it takes for the child to respond to those.
GP: When running experiments with human participants, it is crucial to be mindful of all the ethical implications. Social robots are not invasive tools that could put people’s safety at risk. They do however store personal data with their cameras and microphones. In any case, a researcher needs to be knowledgeable about the many laws and regulations concerning experiments with minors and people with disabilities. And it’s true that they can be quite restrictive compared to those concerning typically developing children or adults. Beside that, participants cannot and should not be treated as mere research subjects. It’s important to obtain their consent (and their parents’ in case of underage participants) after informing them about the aims of the study and the activities they will take part in. When setting up an experiment, therapists should ideally be involved as well to ensure an environment adapted to the needs of autistic children. Finally, the children should be reassured that they are free to quit the experiment at any time, if they feel discomfort, without any negative consequences.
If you want to know more about Dr. Palestra’s work, you can check his page here.
2 Comments
Hi Dr Palestra and readers! If you are interested in this work or doing related work, you may be interested in an upcoming workshop on child-robot interaction at the IDC (Interaction Design and Children) conference, whcih will include representation of projects on robots for autism. See the workshop website for workshop and submission information: https://sites.google.com/view/cri-idc-workshop/
Hello, Alyssa! The conference would be of great interest to our community. Thanks for sharing!