Framework development of human-robot interaction for autism spectrum disorder (ASD)

Abstract

Autism Spectrum Disorder (ASD) is a neurological and developmental disorder that affects one’s ability to understand social cues and to communicate. As it begins early in childhood and can last throughout a person’s life, it is important to have an early intervention and rehabilitation. Previous research has shown that robotic platforms helped and encouraged positive outcome in ASD children rehabilitation. Early interventions through human-robot interaction (HRI) have the potential when used to improve communication and social skills of the children. Currently, the existing robots used in rehabilitation tend to have specific features and appearance and are lacking in ability to extend their capability to do extra tasks in encouraging engagement of ASD children towards learning and interaction. In this research, a new framework has been developed featuring a robot and human interaction modules that includes robot selection, development of interaction modules, and HRI measurement. A commercial robot, Rero has been selected based on its reconfigurable capability, mobile, speech enabled, controllable, programmable, and attractive characteristics. In order to determine which robot configuration is more preferable by the ASD children, a humanoid robot and a pet-like mobile robot have been built and their preliminary responses have been observed. A few interaction modules have been developed and they are consist of imitating actions, following instructions, naming objects, as well as focusing and matching colours. Two sets of questionnaires which consist of task completion (Part 1) and engagement level (Part 2) have been analysed. Observations from the experiment shows that 75% of the children enjoyed the interaction modules and 70% were well engaged with the robots. The engagement was also measured using image-based proxemics behaviour measurement method to see the distance between the robot and the children. From the Sign test, it can be seen that the results from the combination of image-based proxemics behaviour measurement (average distance) and Questionnaire Part 2 is better than results from Questionnaire Part 2 only. Nine ASD children have been selected for this experiment based on certain criteria. From the results, two out of nine subjects have no preference of robots, five of them prefer the mobile robot and another two prefer the humanoid robot.