Cancellation of Unwanted Doppler Radar Sensor Motion Using Empirical Mode Decomposition

Abstract

The operation of microwave Doppler radar for sensing physiological motion signals is heavily compromised under sensor motion. To that end, we investigate the feasibility of applying empirical mode decomposition method in this context, and demonstrate its effectiveness in removing sensor motion artifacts. This method is shown to be effective in canceling unwanted sensor motion with precision sufficient to enable accurate heart rate extraction. Theoretical analysis and simulation results illustrate the potential of the proposed approach for a wide range of frequency separation and amplitude ratios of physiological signals and motion artifacts. Experimental results confirm that separation success is not very sensitive to amplitude ratio. A heart rate is extracted with RMSE within 1 beat per minute even in the presence of mechanical motion and order of magnitude larger in amplitude than that of the heart signal.

Cyclostationary modelling of amplitude and frequency modulated signals in heart and respiration monitoring Doppler radar systems

“…Therefore, this expected value, which has second-order cylostationary characteristics according to (16), is interpreted as a statistical ensemble average [21]. □ We exemplify the above theorems using analysis and experimental results in the following section.…”

“…As seen from (16), the SCF of the F y(t 1 )y * (t 2 ) (x 1 , x 2 ) ; lim signal will have all of its frequencies along the temporal frequency f-axis, because of the non-stationary nature of the signal, and discrete frequencies that correspond to hidden periodicities along the cycle frequency α-axis. According to (16), the spectral lines along the cyclic frequency axis have …”

“…In [12], a methodology is presented to choose the best carrier frequency to enhance the desired heart and respiration components, which is preferred for frequency estimation. Other approaches include adaptive dc calibration [13], blind source separation, passive radio-frequency (RF) tags [14,15] and noise and clutter cancellation [16] in order to isolate multiple subjects and to diminish the effects of clutter and noise for practical applications of biomedical radar [16]. While accurate heart and respiration rates have been extracted for relatively stationary and isolated subjects [17], it is still a main challenge to obtain reliable data in the presence of human random body motion and other moving objects around [18][19][20].…”

Human Vital Signs Detection Methods and Potential Using Radars: A Review

“…Another method used to cancel RBM effect in vital signs acquisition is the radar implementation with self-injection locking [128]. In [129], the empirical mode decomposition (EMD) signal-processing technique was used to remove motion artifacts from the antenna and the subject. Another signal-processing method was used in [130] to remove the RBM effect on human vital signs.…”

“…Furthermore, signal processing techniques can be used to remove unwanted body movement in Doppler radars for vital signs detection. In [129], the empirical mode decomposition (EMD) technique was used to remove motion artifacts from the antenna and the subject. In [130] the authors demonstrated the possibility of measuring vital signs of humans by apply the cyclostatic transformation to Doppler radars in presence of RBM.…”