A Blind Source Separation Framework for Monitoring Heart Beat Rate Using Nanofiber-Based Strain Sensors

A recently developed novel nanofiber-based strain sensor is introduced as a potential alternative to the conventional measurement tools for heart beat rate monitoring. Since the measured signals in real life are often contaminated by certain artifacts, in this paper, to overcome limitations of currently available empirical mode decomposing (EMD) and blind source separation-based methods and recover the buried heart beat signal accurately, we propose a novel blind source separation framework by combining noise-assisted multivariate EMD (NAMEMD) and multiset canonical correlation analysis (MCCA). The proposed method takes advantage of the multivariate data-adaptive nature of the NAMEMD and MCCA, which contributes to accurate extraction of the desired signal. The absolute correlation coefficients (ACCs) between the extracted signal and the original source signal are adopted to evaluate the performance of the proposed method in the simulation study. The average of the ACC yielded by the proposed method is 0.902, which is significantly higher than that by the state-of-the-art approaches. We also examine the proposed method on the nanosensor data collected when the subject performs 11 tasks. It is shown that the proposed method can achieve better performance, especially for preserving the shape of the heart beat signal.