Georgia Institute of Technology
Biomechanical Feature Evaluation for Movement-Quality Classification in Robot-Facilitated Exercise for Parkinson’s Disease
Abstract
dc:description.abstractParkinson’s disease (PD) is partially characterized by motor impairments that significantly affect movement quality, which makes accurate assessment of human motion imperative for rehabilitation efforts and long-term monitoring. Targeted exercise is a commonly prescribed and clinically effective treatment for managing PD symptoms and slowing functional decline. Current therapeutic exercise apparatuses have continually and increasingly relied on robotic assistance to chaperone movement and provide feedback. Separately, these systems can integrate wearable sensors and human pose estimation to capture biomechanical data to inform feedback provided during exercise. However, it is unclear which biomechanical features derived specifically from camera-based human pose estimation signals best correlate to the quality of movement. This thesis presents an interpretable human kinematic model designed to extract biomechanical features by converting IMU-derived joint angles into Cartesian coordinate positions of major human body parts similar to those predicted by 3D human pose estimation Using expert labels of exercise form provided by physical therapists, we evaluate the single-feature logistic regression and decision tree models with features as the input in order to classify correct or incorrect exercise form. The results demonstrate that the time series features of Cartesian joint positions are sufficient inputs to single-dimensional logistic regression models for exercise form classification. This work provides an interpretable, feature-prioritized methodology for understanding the quality of movement assessment in PD.
Degree
thesis:*- Name thesis:degree_name
- Biomedical Engineering, MS
- Department thesis:degree_department
- Biomedical Engineering (Joint GT/Emory Department)
- Grantor
- Georgia Institute of Technology
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Rodriguez, Ethan Alexander
- Advisors dc:contributor.advisor
-
- Hackney, Madeleine
- Gleason, Rudolph L.
- Danziger, Zachary
- Committee member dc:contributor.committeemember
-
- Johnathan Mckay
Rights
- Language dc:language.iso
- English
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/81640