Affect Recognition in Robot Assisted Rehabilitation of Children with Autism Spectrum Disorder

This paper presents a novel affect-sensitive human-robot interaction framework for rehabilitation of children with autism spectrum disorder (ASD). The overall aim is to enable the robot to detect and respond to the affective cues of the children in order to help them explore social interaction dynamics in a gradual and adaptive manner. The first part of the proposed framework, namely the 'affect recognition' module is developed in detail in this paper. Two tasks are designed to elicit the affective states of liking, anxiety, and engagement that are considered important in autism rehabilitation. Affective cues are inferred from psychophysiological analysis that uses subjective reports of the affective states from a therapist, a parent, and the child himself/herself. Comprehensive physiological indices are investigated that may correlate with the affective states of children with ASD. A support vector machines based affect recognizer is designed that yielded reliable prediction with approximately 83% success when using the therapist's reports. This is the first time, to our knowledge, such a human-robot interaction framework for autism rehabilitation is proposed. This is also the first time that the affective states of children with ASD have been experimentally detected via physiology-based affect recognition technique.