|
Project properties |
|
| Title | Extracting Heart Rate from Noisy PPG Signals in Dairy Cows |
| Group | Systems and Synthetic Biology |
| Project type | thesis |
| Credits | 36 |
| Supervisor(s) | Jan Aarts (Adaptation Physiology Group) Edoardo Saccenti (Systems and Synthetic Biology) |
| Examiner(s) | Jan Aarts, Edoardo Saccenti |
| Contact info | robert1.smith@wur.nl |
| Begin date | 2025/06/13 |
| End date | |
| Description | This thesis project will be a computational project. The project will be supervised by Jan Aarts (Adaptation Physiology Group) and Edoardo Saccenti (Systems and Synthetic Biology Group).
At the Adaptation Physiology Group, we study how animals cope with changing environments and stress. One of the key indicators for stress, and overall health and welfare is heartrate (HR). Especially the variance in heartrate, heart rate variability (HRV), has proven to be a good indicator of an animal’s health status. Continuous HR monitoring is animal is technically challenging. While ECG-based heart rate belts are widely used in humans, they are often not suited for animals. An alternative approach is photoplethysmography (PPG), a non-invasive method using light refraction on the skin [1]. However, the signal is very susceptible to motion artefacts, making continuous HR monitoring difficult. The project aims to develop an algorithm to reliably extract HR from the noisy PPG signals collected from dairy cows. A recent paper presented a ML-based approach to reconstruct the heartrate from this signal in children and we would like to adapt this approach such that it can be used on dairy cows [2]. There are several approaches to solve this problem including: • Identify and classify low-noise segments to calculate heartrate from PPG directly. • Train a regression model to estimate HR directly from noisy PPG signals. • Integrate motion data from sensors to inform signal quality or guide the ML model. The dataset includes sensor data collected using a Shimmer sensor system [3]. The available signals include: • PPG • 9-axis motion (accelerometer, gyroscope, magnetometer) • Temperature, pressure • A ground truth heart rate signal obtained from a reference HR belt system References related to project: 1. https://doi.org/10.1016/j.bpa.2014.08.006 2. https://doi.org/10.3390/bios13070718 3. https://www.shimmersensing.com/product/optical-pulse-probe/ |
| Used skills | |
| Requirements | Required competences for this project include:
• Experience with data analysis and signal processing in Python • Familiarity with classification and regression techniques • Basic knowledge on time series modelling is useful but not required |