AI-Assisted Sensing and Support for Caregiver–Child Interactions in Autism
Project Description
Caregivers play an important role in supporting autistic children during everyday activities, communication, learning, and emotionally challenging situations. However, understanding a child’s needs and deciding how to respond can be difficult, particularly when behavioral and physiological cues are subtle, rapidly changing, or different across individuals.
Recent advances in mobile sensing, wearable devices, computer vision, and artificial intelligence create new opportunities to support caregivers in understanding caregiver–child interactions. Rather than diagnosing autism or replacing professional care, such technologies may help caregivers recognize meaningful interaction patterns, review important moments, and receive timely, context-appropriate support.
This project will explore human-centered sensing and AI methods for understanding and supporting caregiver–child interactions. Possible research components include collecting and analyzing behavioral, physiological, motion, or contextual information; developing computational methods for identifying changes in engagement or interaction state; and designing caregiver-facing interfaces that provide summaries, feedback, or guidance.
Depending on the applicant’s interests and the progress of the project, the work may focus on sensing-system development, multimodal data analysis, machine learning, interface design, or user evaluation. The exact sensing modalities, target scenarios, computational methods, and study procedures will be refined during the project.
The long-term objective is to develop practical, respectful, and low-burden technologies that help caregivers better understand and support autistic children in everyday settings.
Recent advances in mobile sensing, wearable devices, computer vision, and artificial intelligence create new opportunities to support caregivers in understanding caregiver–child interactions. Rather than diagnosing autism or replacing professional care, such technologies may help caregivers recognize meaningful interaction patterns, review important moments, and receive timely, context-appropriate support.
This project will explore human-centered sensing and AI methods for understanding and supporting caregiver–child interactions. Possible research components include collecting and analyzing behavioral, physiological, motion, or contextual information; developing computational methods for identifying changes in engagement or interaction state; and designing caregiver-facing interfaces that provide summaries, feedback, or guidance.
Depending on the applicant’s interests and the progress of the project, the work may focus on sensing-system development, multimodal data analysis, machine learning, interface design, or user evaluation. The exact sensing modalities, target scenarios, computational methods, and study procedures will be refined during the project.
The long-term objective is to develop practical, respectful, and low-burden technologies that help caregivers better understand and support autistic children in everyday settings.
Supervisor
SHAO, Qijia
Quota
2
Course type
UROP1000
UROP1100
UROP2100
UROP3100
UROP3200
UROP4100
Applicant's Roles
The applicant will participate in different stages of the research project. Possible responsibilities include:
Reviewing related work in autism support, caregiver–child interaction, human sensing, and human–AI interaction.
Supporting the identification of meaningful caregiver–child interaction scenarios and research questions.
Assisting with the design and development of sensing or data-collection prototypes.
Supporting pilot studies and human-participant data collection, annotation, organization, and quality checking.
Processing behavioral, physiological, motion, video, or other multimodal data.
Developing and evaluating computational methods for recognizing interaction patterns or user states.
Contributing to the design of caregiver-facing summaries, feedback, or guidance interfaces.
Evaluating the usability, usefulness, privacy, and acceptability of the proposed system.
Contributing to research documentation, presentations, and potential paper preparation.
The applicant’s specific responsibilities may be adjusted according to their background, interests, and the progress of the project. Students may place greater emphasis on sensing, artificial intelligence, human–computer interaction, or experimental evaluation.
Reviewing related work in autism support, caregiver–child interaction, human sensing, and human–AI interaction.
Supporting the identification of meaningful caregiver–child interaction scenarios and research questions.
Assisting with the design and development of sensing or data-collection prototypes.
Supporting pilot studies and human-participant data collection, annotation, organization, and quality checking.
Processing behavioral, physiological, motion, video, or other multimodal data.
Developing and evaluating computational methods for recognizing interaction patterns or user states.
Contributing to the design of caregiver-facing summaries, feedback, or guidance interfaces.
Evaluating the usability, usefulness, privacy, and acceptability of the proposed system.
Contributing to research documentation, presentations, and potential paper preparation.
The applicant’s specific responsibilities may be adjusted according to their background, interests, and the progress of the project. Students may place greater emphasis on sensing, artificial intelligence, human–computer interaction, or experimental evaluation.
Applicant's Learning Objectives
Develop an understanding of human-centered technology for autism and caregiver support.
Gain hands-on experience in mobile, wearable, or multimodal sensing.
Learn fundamental methods for processing behavioral and physiological data.
Apply machine-learning techniques to real-world human-interaction data.
Understand how to design AI systems that provide context-sensitive and caregiver-centered support.
Gain experience in experimental design, human-participant research, data privacy, and responsible AI.
Learn to evaluate both technical performance and human-centered outcomes.
Improve research communication, interdisciplinary collaboration, technical documentation, and presentation skills.
Gain hands-on experience in mobile, wearable, or multimodal sensing.
Learn fundamental methods for processing behavioral and physiological data.
Apply machine-learning techniques to real-world human-interaction data.
Understand how to design AI systems that provide context-sensitive and caregiver-centered support.
Gain experience in experimental design, human-participant research, data privacy, and responsible AI.
Learn to evaluate both technical performance and human-centered outcomes.
Improve research communication, interdisciplinary collaboration, technical documentation, and presentation skills.
Complexity of the project
Moderate