A Robotic Framework to Overcome Sensory Overload in Children with Autism Spectrum Disorder: A Pilot Study

INTRODUCTION Experiences of sensory overloads are reported to be central to the autistic experience. A higher prevalence of such unusual sensory responses is seen in children on the autism spectrum, when compared to their typically developing and developmentally delayed counterparts. This work focuses on a novel framework designed to teach children with Autism Spectrum Disorders (ASD) how to react to a variety of sensory stimuli. We hypothesize that observing the robot’s reactions to stimuli in socially acceptable ways will enable the children to mold their own behaviors accordingly. THE FRAMEWORK Two different robotic platforms are used: a mobile robot with a character display, Romo, for facial expressions, and a small humanoid robot, Mini, for emotional expression through gestures and movements. We first introduce the child to Romo’s expressions through an Emotion Game, implemented as a computer game, whereby children react to each of Romo’s emotions by selecting the emotion it makes them feel. The next set up consists of five sensory stations, one for each of the visual, auditory, olfactory, gustatory and tactile senses. The robotic agent navigates them one by one, and responds to the different stimuli presented to it. The stimuli are chosen to emulate scenarios, both negative and positive, encountered by children in their everyday lives, some of which may possibly be sources of sensory overload to these children. RGB-D sensors are used to record and quantify children’s movements. All sessions are video-recorded. PILOT STUDY Using this framework, we conducted a small pilot study consisting of eight participants, three children with ASD (two played the emotion game only) and five neurotypical children. Each session lasted about 15 minutes. The purpose of this preliminary study was to identify a general response to our framework, and to verify if the children’s responses meet our expectations. Our next step is to conduct a more long-term user study, involving a much larger number of participants, to improve the reliability of the obtained results. RESULTS Most participants clearly displayed high levels of engagement in the interaction and were able to accurately interpret the robot’s behaviors to a large extent. One child with ASD, even though he did not focus entirely on the activity at hand, imitated the robot’s behaviors voluntarily, and was enthusiastic about learning how to control the robots himself. RGB-D sensor data shows meaningful patterns of short-term effects, together with qualitative analysis of the video recordings