Author Date

2025-12-15

Degree Name

BS

Department

Mathematics

College

Physical and Mathematical Sciences

Defense Date

2025-11-24

Publication Date

2025-12-15

First Faculty Advisor

Tyler Jarvis

First Faculty Reader

Rob Davis

Honors Coordinator

Davi Obata

Keywords

windkessel, photoplethysmography, ppg, cardiovascular disease

Abstract

Features of cardiovascular health such as blood pressure, cardiac output, arterial compliance, and blood viscosity are important in evaluating heart health and disease risk, but are currently largely measured invasively and therefore only in critical settings. Here we explore the potential for Windkessel models to extract cardiovascular features from photoplethysmography (PPG) data in an attempt to evaluate heart health non-invasively. We show the capacity of a constrained Newton's method optimization scheme to estimate parameters and explore the tractability of the resulting parameters. While Newton's method is able to produce parameters that closely approximate PPG data, we found that using a Windkessel model to produce peripheral PPG from aortic PPG is an unidentifiable problem, and the parameters are therefore uninterpretable. Further, any approach to solving the inverse problem of pulse generation via the Windkessel model is intractable.

Download

Share

COinS