Enhancing social skills group intervention for children with autism spectrum disorder through socially assistive robotics

• Main Author: Wong, Alvin Hong Yee
• Other Authors: Zhong Zhaowei
• Format: Theses and Dissertations
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering::Mechanical engineering::Robots; DRNTU::Social sciences::Sociology::Social behavior; DRNTU::Engineering::Industrial engineering::Human factors engineering
• Online Access: http://hdl.handle.net/10356/69016
• Institution: Nanyang Technological University

Children characterised with Autism Spectrum Disorder (ASD) would be somewhat reserved, deficient in joint attention and socially withdrawn. Although no remedy had been identified for ASD at the present moment, early intervention could arrest its degradation. As ASD could manifest itself in many diverse forms, there would be many forms of intervention. Studies had demonstrated the veracity of social robots in complementing the human therapist. With a robot, investigators could isolate particular responses, and therefore could create a novel way of treatment. Earlier research showed social robots could possibly interact with children with ASD in three distinct roles. Firstly, as a playmate, sustaining a child's interest and attracting a child to be taught social skills. Next, to be a social moderator that could promote social norms learning. Lastly, to be a mediator that could foster learning of social skills in a natural environment. Apart from difficulty in establishing eye contact, children with ASD would experience difficulty developing joint attention. However there was no published research on what type of verbal and non-verbal articulation could affect the level of joint attention. Hence this thesis examined the use of robots in augmenting the learning outcomes of social skills and improving levels of engagement in a social skills group based intervention with children with ASD. The Theory of Mind (ToM) was utilised to create a model (including the Emotion Recognition & Expression (EREx) framework) for designing the robot’s behaviour to obtain reactions from the children to the robot’s actions. The ToM expounded that one could be aware of one’s own mental states as well as those of others, which might be different from one’s own. Using Computer as Social Actors (CASA) Theory as a basis, this research found that the key importance was to design robots that could comprehend children with ASD and children’s actions. This work also built upon Model of Communication and Framework on System Image. A new framework was proposed whereby practitioners could have an intended social skill, to communicate to users via an image of a robot. Social skills were presented via verbal and/or non-verbal communication from the robot. The users would then perceive these “images” through verbal speech and emotions through intended non-verbal communication. While this might seem intuitive to a typically developing child, children with ASD would have difficulties in this social behaviour as illustrated by ToM. Scientific contributions in this research included establishing a novel framework to enhance social skills group intervention for children with ASD through socially assistive robotics. Using this framework, four studies were conducted to evaluate the proposed robot training as a potential intervention for children with ASD. The first preliminary study (N=60) was carried out to investigate and understand user acceptance, attitude and response towards the CuDDler robot. A 2nd preliminary study (N=102) was conducted to see whether users understood the robot’s emotional acts. The outcome of two preliminary studies then led to a consequential Study I (N=28) that investigated how children with ASD engaged joint attention learning with a robot. Finally, Study II (N=48) looked at how effective the robot could impart social skills to a child with ASD. The robot training method had produced positive outcome showing improvements in certain social skills for children with ASD. The contribution of this study was three-fold. Firstly, this study expounded the idea of ToM that could be utilised for constructing social robots that correspond with people and teach social skills. Secondly, significant improvements in turn-taking skills and more importantly, joint attention was identified during the robot training. Thirdly, this research identified faster improvements in level of engagement through scientific outcome measures. The thesis had presented empirical evidence of the robot’s key role in providing an effective social skills training to children with ASD, focusing on joint attention, turn-taking and maintaining eye contact. Future studies might take this as a basis on which to expand on, to examine this method in more detail, test it on an even larger sample size and broader spectrum of children with ASD over a longer period of time and also possibly design and test variations of this method to train and improve other cognitive mechanisms.