Wearable sensors for smart abnormal heart rate detection during skiing

Abstract

Abnormal heart rate detection has widely used in the evaluation, diagnosis and prediction of heart-related diseases and autonomic nerve-related diseases. In this paper, we use the ballistocardiogram (BCG) signal to analyze abnormal heart rate. First, we collect the BCG signals from wearable device during skiing. Second, we use the empirical model decomposition (EMD) and Hilbert transform to remove the noises in collected BCG signals. Third, the denoised BCG signals are used to analyze the heart rate variability (HRV) which is widely used in medical diagnosis. The experimental results show that the HRV analysis based on BCG signal and ECG signal has no significant difference. Lastly, the HRV features are input into a local outlier factor (LOF) to identify abnormal heart rate during skiing.