Online Smartwatch System for Emotional Regulation in People with Autism Spectrum Disorder

No Thumbnail Available
Issue Date
Journal Title
Journal ISSN
Volume Title
One of the symptoms of Autism Spectrum Disorder (ASD) is emotional dysregulation. Even in cases of high-functioning autism, the tendency to- wards intense emotional reactions to speci c situations can greatly hinder a person's quality of life. This project aims to develop a personalized wear- able system that learns to detect the person's unique stressful responses, which in turn can aid the user to better recognize triggering moments and learn appropriate coping skills. The system is developed and tested on every day out-of-lab use, where normal constraints on data gathering cannot be applied. The rst software prototype is built without a ma- chine learning component for the purposes of data collection and feature extraction. For the nal version several online learning algorithms are compared and a 1-layer neural network with a Stochastic Gradient Des- cent optimizer is selected to be used in the study. The experiment aims to learn whether the chosen model can adapt to user feedback and learn to accurately classify stressful moments as de ned by each individual. In turn, the project also looks at the participants are helped by the system to recognize these moments. The results do not provide conclusive answers as the experiences of the participants varies. However, steps for improving the software in future work are outlined in the closing discussion.
Faculteit der Sociale Wetenschappen