Title: Robot-led Rehabilitative Exercise for People with Parkinson's Disease

Date: Tuesday, April 14th, 2026

Time: 3PM - 5PM ET

Location: 57 Executive Park S, room 108

Virtual link: https://gatech.zoom.us/j/94672782852

Committee:

Dr. Madeleine E. Hackney (Advisor) - School of Medicine, Department of Geriatrics and Gerontology, Emory University & Wallace H. Coulter Department of Biomedical Engineering, Emory University and Georgia Institute of Technology & Research Health Scientist, Atlanta VA Health Care System

Dr. Lena Ting - Wallace H. Coulter Department of Biomedical Engineering, Emory University & Georgia Institute of Technology

Dr. Gregory Sawicki - George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology

Dr. Hyeokhyen Kwon - School of Medicine, Department of Biomedical Informatics, Emory University

Dr. Charles C. Kemp - Chief Technology Officer, Hello Robot Inc.

Abstract:

Regular exercise is a key component of rehabilitative treatment for Parkinson’s disease (PD), yet people with PD (PWP) often struggle to adhere to prescribed exercise protocols outside of supervised clinical settings. Technology-based interventions have been proposed to address this gap, but existing systems do not fully replicate the guidance and support provided during in-person therapy. This dissertation presents the design of the Zesty Exercise System for Therapeutic Engagement (ZEST-E), a robot-led rehabilitative exercise system designed to guide PWP through repetitive, large-amplitude stretching exercises. Through two user studies involving PWP and exercise specialists, i.e., physical therapists, this work demonstrates the feasibility of robot-led rehabilitative exercise and identifies key facilitators and barriers to deploying such systems beyond the clinic. In addition, this dissertation characterizes how PWP perform exercises with ZEST-E with a focus on adherence to prescribed movement form. These data are used to develop an automated human activity recognition pipeline that classifies exercise form correctness using IMU-derived features. Together, these contributions support the feasibility of robot-led rehabilitative stretching exercise for PWP and provide initial methods for assessing exercise performance in this context.