Developing an effective monitoring system using sensors and artificial intelligence for sensory regulation of children with autism spectrum disorders

Abstract

In the field of Autism Spectrum Disorder (ASD) intervention, there has been a growing need for technology-based methods to address atypical sensory responses, a core symptom experienced by most children with ASD. Atypical sensory responses lead to their difficulties in self-regulation in daily life. They may have difficulty paying attention or recovering from anxiety. Sensing technologies and artificial intelligence (AI) in collaboration represent a promising tool because they not only enable real-time monitoring of the sensory responses, but can also produce useful intervention strategies for assisting children with ASD. The aim of this research is to explore how to develop an effective and acceptable intelligent system, using reliable sensor and AI techniques, to facilitate sensory regulation of children with ASD. A monitoring system named Roomie, has been proposed and developed as a tool to explore the research questions. The research followed a user-centred principle and iterative process, which means that Roomie was developed with the help of ASD specialists and user groups, and had been constantly refined. Multiple sensors were used to collect environmental data and physiological data, in order to obtain a comprehensive understanding of a child’s sensory responses in relation to their environment. A standardised sensory profiling tool was integrated in the system to obtain information about a child’s sensory processing pattern. Machine learning algorithms were used to extract and analyse sensory-related data to detect the child’s attention and stress levels. A fuzzy logic algorithm was employed to stimulate the strategy-making process of an ASD specialist based on the detected environmental stressors and abnormal states. Key modules such as data processing and feedback generating were integrated in a smartphone-based application, which make the system easier for children with ASD and their caregivers to access. The entire system has been tested in a series of evaluation sessions in a real-life setting. Overall, the results presented in this thesis suggest that the proposed sensor and AI-enabled system can effectively address atypical sensory responses in children with ASD. At the end of the thesis, discussion on the further improvement and wider application of the system has been made. The work presented in this thesis has provided a solid foundation for future studies in which the proposed system and development framework can be used for creating a smart health home to implement the environmental control and sensory regulation strategies automatically without a continuous involvement of a human assistant.

Date of Award	Jul 2024
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Anthony Graham Bellotti (Supervisor), Adam Rushworth (Supervisor), Sarah Dauncey (Supervisor) & Prapa Rattadilok (Supervisor)