Heart rate monitoring using PPG signals

Abstract

This bachelor’s thesis discusses the idea and implementation of real-time heart rate monitoring using photo- plethysmography (PPG) signals. PPG signals measured from the wrist are often subjected to distortion and noise. Signal processing techniques that tackle these issues were investigated and implemented. The algo- rithm of choice was JOSS. JOSS consists of sparse signal reconstruction, spectral subtraction and spectral peak tracking. The combination of these techniques is what results in a robust heart rate tracking algorithm. An improvement, in terms of computing power, was made by analyzing general properties of PPG signals. The outcome is a new framework named SMART, which combines the HR tracking power of a robust HR monitor- ing system, like JOSS, with the speed of faster HR monitoring methods. The overall result is a fast algorithm that computes the heart rate with an average absolute error of 1.42. This result shows that PPG based heart rate monitoring on wearable devices has great potential for fitness and medical purposes.