Multichannel Bed Based Ballistocardiography Heart Rate Estimation using Continuous Wavelet Transforms and Autocorrelation

Ismail Elnaggar, Tero Hurnanen, Jonas Sandelin, Olli Lahdenoja, Antti Airola, Matti Kaisti, Tero Koivisto
University of Turku


Bed Based Ballistocardiography (BCG) has recently garnered more attention due to increased interest in non-invasive cardiac monitoring specifically related to remote patient monitoring. BCG is a prime candidate for at home and nighttime monitoring especially in the growing elderly population because co-operation from the user is not required to be able to record signals. One issue with BCG is that the signal quality can vary from person to person based on factors such as body position and motion artifacts, making it challenging to identify individual heartbeats. The goal of this study is to evaluate a multichannel bed sensor setup to address these issues.

A rule based algorithm which considers all eight available BCG channels simultaneously from a given time epoch was developed using continuous wavelet transform (CWT) to extract the localized time-frequency representation of each epoch and then an averaging method was applied across the different scales of the CWT to produce a flattened array. Autocorrelation was then applied to this array to produce a heart rate estimate. This method does not require identification of individual heart beats to accurately estimate heart rate and does not require training data.

The data is taken from an open dataset published in 2021. This dataset contains time aligned ECG and BCG from 40 participants (17 males, mean age 34 +/- 15) totaling 4.5 hours of BCG data. There are eight-channels of BCG signals from different positions on the bed.

When applied to 4 second time epochs, this model produces an average mean absolute error (MAE) of 1.09 bpm when compared to heart rate derived from ECG. The minimum MAE across all 40 subjects was 0.35 bpm and the maximum MAE was 3.55 bpm. This method produces competitive results without the need for annotated training data, which can be challenging to collect.