McMaster University
Multimodal Wearable Systems Towards Continuous Monitoring of Mobility and Physiological Health
Abstract
dc:description.abstractContinuous and objective assessment of human mobility and physiological health is essential for the early detection, monitoring, and management of age-related functional decline and chronic conditions. However, conventional clinical assessment tools are often constrained by high cost, short observation windows, and limited ecological validity. To address these limitations, robust, low-cost, and interpretable wearable sensing frameworks for holistic mobility and health assessment in real-world environments are developed and presented in this thesis. A magnetometer-free inertial measurement unit (IMU)-based framework is established to jointly estimate lower-limb kinematics, reconstruct drift-corrected spatial trajectories, and extract comprehensive gait parameters during ambulatory walking. The framework integrates adaptive filtering and stride-level drift compensation and is validated against optical reference systems, demonstrating reliable estimation of joint kinematics and spatial gait measures under real-world conditions. Building on this framework, wearable-derived gait and joint features are characterized using machine-learning methods to identify mobility patterns associated with age, body mass index, and sex. To enhance clinical interpretability, a simple Human Gait Index is formulated by integrating key spatiotemporal and kinematic parameters into a compact and interpretable measure of overall gait quality. In parallel, a low-cost, chest-mounted multimodal wearable system is developed and validated for continuous physiological monitoring. The system integrates synchronized electrocardiography, photoplethysmography, inertial, temperature, and acoustic sensing to estimate cardiovascular, respiratory, and infection-related parameters during daily activities. Finally, real-world remote monitoring is applied in a cohort of older adults with and without chronic low back pain, demonstrating the feasibility of long-term wearable assessment and showing that machine-learning models applied to activity-derived features can distinguish pain-related mobility patterns that are not observable through conventional clinic-based evaluations. Overall, the work presented in this thesis advances wearable sensing methodologies that bridge biomechanics, physiology, and data-driven analytics, enabling scalable, interpretable, and ecologically valid monitoring to support tele-health applications and personalized care.
Degree
thesis:*- Department dc:contributor.department
- Electrical and Computer Engineering
- Grantor dc:publisher
- McMaster University
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Faisal, Abu Ilius
- Advisor dc:contributor.advisor
-
- Deen, M Jamal
Subjects
dc:subject × 10Rights
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Identifier URI
- https://doi.org/10.71548/5