Automatic heart rate detection from FBG sensors using sensor fusion and enhanced empirical mode decomposition.

Cardiovascular diseases are the world's top leading causes of death. Real time monitoring of patients who have cardiovascular abnormalities can provide comprehensive and preventative health care. We investigate the role of the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and sensor fusion for automatic heart rate detection from a mat with embedded Fiber Bragg Grating (FBG) sensor arrays. The fusion process is performed in the time domain by averaging the readings of the sensors for each sensor array. Subsequently, the CEEMDAN is applied to obtain the interbeat intervals. Experiments are performed with 10 human subjects (males and females) lying on two different positions on a bed for a period of 20 minutes. The overall system performance is assessed against the reference ECG signals. The average and standard deviation of the mean relative absolute error are 0.049, 0.019 and 0.047, 0.038 for fused and best sensors respectively. Sensor fusion together with CEEMDAN proved to be robust against motion artifacts caused by body movements.