S Jerritta
Emotion regulation in people with Autism Spectrum Disorder (ASD) is challenging for family, caretakers and people around them due to extreme behavior patterns such as aggression, self-injury, defiance and outbursts. Understanding the underlying emotional states of the children can help care takers, parents, teachers and other concerned persons to intervene and formulate personalized preventive and reactive strategies. A number of wearable devices like smart watches and smart belts are used for monitoring the health and related parameters. Researchers in Human Computer Interaction (HCI) have used these devices to identify the hidden emotional state of the user by acquiring physiological data (Electrocardiogram (ECG), Electromyogram (EMG) etc.,) as well as visible behavioral information (posture, gesture, activity levels etc.,) to predict the emotional state of the user. This research aims to develop an AI device that would predict the internal state of children suffering from Autism Spectrum Disorder (ASD) using the Heart Rate Variability (HRV) signals derived Electrocardiogram signals (ECG). Data is collected corresponding to the positive and negative valance of children with ASD and controls. Results indicate higher order statistical features to significantly demarcate the positive and negative states of children with ASD. Validating the results and embedding the algorithm into wearable AI devices can help in identifying the internal component of the child’s behavior and provide personalized care.