University of British Columbia
Secure and scalable human motion understanding using machine learning-powered smart sensing textile
Abstract
dc:descriptionThis dissertation explores how advanced wearable technology can enhance our ability to understand and track human movement. It focuses on the development of smart textiles, such as gloves, arm sleeves, and knee sleeves, which incorporate sensor yarns and motion sensors to accurately monitor body movements in real time. These textiles are designed to be durable and practical for everyday use, capable of withstanding repeated wear and washing. The study investigates six distinct applications, cross-validated across different subjects and sessions, demonstrating robust accuracy and the potential of smart textiles in fields like virtual reality, healthcare, robotics, and human-computer interaction. A key aspect of this research is the development of a system to support post-stroke rehabilitation. The system utilizes smart gloves to assess patients’ hand and arm movements, providing scores based on range of motion, muscle strength, and speed. The knee sleeve also evaluates body movements, demonstrating its potential for clinical use. In addition to these specific applications, the research examines broader challenges related to smart apparel and their integration with IoT devices. We addressed two key areas: the scalability of these systems and the sharing of data between IoT devices. To efficiently manage the large volume of data generated by wearable devices, the dissertation introduces SemiPFL, a machine learning approach that enables secure, personalized learning even with limited labeled data. This framework is designed to handle diverse data types effectively, making it suitable for large-scale health monitoring and other applications. Furthermore, the research investigates the generation of realistic human movements using machine learning models that adhere to physical laws, aiming to prevent common artifacts like unnatural foot motions. The goal is to develop customizable models capable of generating natural-looking movements across various scenarios. This work represents significant advancements in wearable technology and the understanding of human movement, with promising applications in remote healthcare, personalized rehabilitation, and human-computer interaction.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy - PhD
- Level thesis:degree_level
- doctoral
- Discipline thesis:degree_discipline
- Electrical and Computer Engineering
- Grantor dc:publisher
- University of British Columbia
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Tashakori, Arvin
- Contributors dc:contributor
-
- Servati, Peyman
- Wang, Z. Jane
Rights
dc:rights- Statement dc:rights
-
- Attribution-NonCommercial-NoDerivatives 4.0 International
- Language dc:language
- eng
Identifiers
dc:identifier.*- Handle dc:identifier
- http://hdl.handle.net/2429/90894