Accurate Fast Heartrate Detection Based on Fourier Bessel Series Expansion During Radar-based Sleep Monitoring

Continuous fast heart rate (HR) detection can provide heart rate variability (HRV), which indicates the autonomous activity system and has been found to change during different sleep stages. However, the accurate fast HR measurement is still challenging due to the frequency resolution limitation and high algorithm complexity of existing methods. In this paper, a novel respiration-reduced Fourier Bessel series expansion (FBSE) technique is proposed to realize fast detection of HR without filtering process using short-time (less than 5s) window length. It is theoretically illustrated that the proposed method has better spectrum resolution. The simulation results also show the proposed method has accurate spectrum representation of heart motion signal. With a custom-designed 24GHz Doppler radar, the overnight sleep experiment was carried out under the clinical standard. The results show the HRV obtained by proposed technique has the good correspondence with the polysomnography signal and has the potential on the future automated HRV-based sleep stage classification.