Towards adaptive and personalized robotic therapy for children with Autism Spectrum Disorder

Abstract

The use of robots for Autism Spectrum Disorder (ASD) research has been witnessed in all of its stages: diagnosis and therapy. This review aims to focus on summarizing the literature specific to autism therapy using robots which consists of two parts: 'activities involved in therapy' and 'perceiving the engagement levels during therapy'. Physiological signals and eye gaze have been used as the first approach to estimate the engagement levels based on the proposed perspective-taking and joint attention activities outlined in this review. This approach will be adopted to reduce multi-modal channels of perceived engagement, which have been found to increase latency as seen from previous literature. In the end, future directions of research based on the limitations of the existing literature have been proposed.