Date of Award
2026
Degree Type
Dissertation
Degree Name
Doctor of Philosophy in Electrical Engineering
Department
Electrical, Computer, and Biomedical Engineering
First Advisor
Kunal Mankodiya
Abstract
A wearable health monitor is a body-worn device, a watch, ring, garment, armband, or patch, that records physiology while a person lives outside the clinic. Its promise is straightforward: many clinically important changes happen over days or weeks, instead of the few minutes clinician gets with the patient during a scheduled visit. Wearable devices are now common enough to screen large populations for atrial fibrillation, estimate sleep state, and track basic cardiovascular markers. Yet the devices people are willing to wear still measure only a narrow set of signals, while many measurements clinicians use remain tied to equipment that is too bulky, intrusive, or short-term for daily life.
This dissertation argues that this gap cannot be closed by miniaturization alone. A wearable that recovers clinical-grade measurements in everyday use has to be designed around three linked questions: what biosignal can be reasonably sensed in a given site, what information can be computed from the biosignal that survives, and what wearable form factor a person can realistically wear. The body site a patient accepts is often not the site where the signal is strongest. The algorithm that recovers the degraded signal has to fit the device’s power and memory budget. A device that is uncomfortable, conspicuous, or difficult to put on will not collect enough data to matter clinically. I test this argument by designing and evaluating wearable systems across opioid use disorder, continuous cardiac hemodynamics, and Parkinson’s disease.
In opioid use disorder, stigma made concealment a design requirement and pushed ECG sensing to the upper arm, where the cardiac signal is much weaker than on the chest. I first characterized dry electrode impedance under realistic contact force and perspiration, then built a custom arm-based acquisition platform and an edge-AI R-peak detector that reduced heart-rate error from motion-corrupted arm ECG. I also developed a cardiovascular digital twin that models each patient’s autonomic response to buprenorphine and estimates individual drug-onset time. In continuous hemodynamics, I moved impedance-cardiography stroke-volume sensing from the standard neck-and-chest array to a chest patch, tested the resulting geometries against an FDA-cleared reference sensor, and recalibrated the governing equation for the new electrode placement. In Parkinson’s disease, the limiting question was simpler and more human: could the participant put the device on without help? I redesigned the sensing glove around independent donning and doffing first, then integrated the movement sensors.
Together, the studies show that clinical wearable design fails when the sensor, algorithm, and enclosure are treated as independent handoffs. Each part constrains the others. The work also exposes a limit that no single wearable can avoid: one body site cannot provide every signal a clinician may want. The dissertation therefore closes by arguing for coordinated wearable networks, such as a ring, armband, and chest patch working together, rather than one ideal device expected to balance every clinical and human constraint on its own.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Recommended Citation
Ravichandran, Vignesh, "DESIGN OF SMART WEARABLES: BALANCING HUMAN FACTORS AND BIOSIGNAL INTEGRITY" (2026). Open Access Dissertations. Paper 4603.
https://digitalcommons.uri.edu/oa_diss/4603
Included in
Biomedical Engineering and Bioengineering Commons, Electrical and Computer Engineering Commons