Design and evaluation of a wearable gait rehabilitation system for stroke patients

Thesis Abstract

Stroke remains a leading cause of long-term disability worldwide, with impaired gait being a predominant obstacle to functional independence among survivors. Current gait rehabilitation approaches often face limitations such as inadequate real-time feedback, limited accessibility, and poor patient compliance, thereby necessitating more effective, tailored interventions that can be conveniently administered in both clinical and home settings. This study aims to design, develop, and evaluate a novel wearable gait rehabilitation system tailored for stroke patients, with the overarching objective of enhancing gait recovery outcomes through personalized, device-assisted therapy. The specific objectives include (1) conceptualizing and designing a lightweight, ergonomic wearable device integrating inertial measurement units (IMUs), wearable sensors, and biofeedback mechanisms; (2) developing an adaptive gait training algorithm grounded in motor learning theories and principles of neuroplasticity; (3) implementing a prototype of the device and conducting preliminary usability testing; and (4) evaluating the effectiveness of the system in improving gait parameters among stroke patients through a mixed-methods clinical trial. The research adopts a quasi-experimental, pretest-posttest design involving 60 stroke patients randomly assigned to an experimental group receiving wearable device-assisted gait training and a control group receiving conventional physiotherapy. The population comprises adult stroke patients within the first six months post-event, recruited from a tertiary rehabilitation hospital. Data collection instruments include standardized gait assessment tools such as the 10-Meter Walk Test (10MWT), the Timed Up and Go (TUG) test, and patient-reported outcome measures (PROMs) like the Stroke Impact Scale (SIS). Additionally, qualitative interviews explore user experience and device acceptability. The validity and reliability of instruments are established through pilot testing and adherence to established psychometric standards. Quantitative data will be analyzed using paired t-tests, repeated measures ANOVA, and regression analysis to assess changes in gait parameters, while qualitative data will be thematically analyzed to understand user perceptions. It is anticipated that the wearable system will significantly improve gait speed, stability, and functional mobility compared to conventional therapy alone, with patients demonstrating higher engagement and adherence to the training protocol. The analytic results expect to reveal statistically significant differences in gait performance metrics with large effect sizes, indicating the device's efficacy in facilitating motor recovery. Furthermore, thematic analysis is expected to identify key factors influencing user acceptance and motivation. The study's findings will contribute empirical evidence supporting the integration of wearable, biofeedback-enabled technology into stroke rehabilitation programs, emphasizing the importance of personalized, feedback-driven interventions grounded in motor learning and neuroplasticity theories such as Schmidt's Schema Theory and the Dynamic Systems Theory. This research advances knowledge by providing a comprehensive framework for designing and empirically validating wearable gait rehabilitation systems tailored for stroke survivors. It offers practical insights into device usability, user engagement, and therapeutic effectiveness. Based on the results, recommendations will be made regarding the commercialization potential, integration into existing clinical workflows, and avenues for future research, including longitudinal studies to assess long-term benefits and scalability across diverse patient populations. The findings aim to inform clinicians, engineers, and policymakers about the viability of wearable assistive technologies in enhancing post-stroke gait recovery and contributing to scalable, patient-centered rehabilitation paradigms.