PPG Heart Rate Detection in the Presence of Motion Artifacts

Abstract

Peripheral circulation can elicit a lot of relevant diagnostic information like heart rate and blood oxygenation level without the need of any invasive measurements. Photoplethysmographic (PPG) signals are obtained by such non-invasive measurements using pulse oximeters. PPG signals, although non-invasive, come with some inherent problems. In a nonhospital environment, like when using a wearable type of sensor, a measured PPG signal predominantly suffers from motion artifacts. Ambient light conditions, temperature, and respiratory artifacts are a few other noise sources that affect the PPG signals when trying to measure heart rates. Most motion artifacts lie in the same frequency range as that of the required noise free signal. So, simple filtering is unlikely to work. This work explores adaptive filtering techniques that are commonly used for noise removal. The current work also proposes to use a popular active noise cancellation technique combined with adaptive filtering and artificial neural networks to minimize the motion artifacts. Furthermore, the work proposes a wrapper algorithm that covers the deficiency of the other techniques. Finally, this work employs a smart peak identification technique to measure reliable heart rates from the MA reduced signals.