Pietro Chierico

PhD Student
Electronic Engineering Department
Universitat Politècnica de València (Spain)

Research line: Biosignal Analysis
PhD Supervisor: José J. Rieta

Doctoral Thesis

Development of Methods for Quality Assessment in PPG Recordings from Wearable Devices

This doctoral research focuses on automatically assessing the quality of photoplethysmography (PPG) recordings from wearable devices. Motion, poor sensor contact, ambient light, and other factors can distort these optical signals and make physiological measurements less reliable. The goal is to develop objective methods that assign recordings to different quality levels, helping subsequent analyses distinguish usable signals from segments that require caution or exclusion.

The research plan combines signal processing with machine learning and deep learning. Quality indicators will describe features such as signal-to-noise ratio and consistency of pulse shape, while accelerometer recordings will help identify motion-related disturbances. Feature-based classifiers and neural networks will be evaluated for their ability to distinguish high-, medium-, and low-quality signals, including unusable recordings.

The proposed methods will be tested on data collected in controlled laboratory settings and during everyday activities, using reference signals and expert annotations to evaluate classification performance. A further objective is to reduce processing time and computational demands so that quality assessment can run in real time on wearable devices. The intended outcome is more dependable selection of PPG data for physiological monitoring. This description summarizes the doctoral research plan and its objectives.