Table 3.

Performance comparison of vital sign detection algorithms for non-stationary subjects.

Vital Sign Assessed	Vital Sign Algorithms	Research Articles	Experimental Setup Range/Subjects/Reference Measurement Method	Results
HR, RR	Fast Fourier transform (FFT), Autocorrelation	[58]	1–2 m/5 humans /ECG	RMSE for RR: 0.006 HR: 0.372
HR, RR	Wavelet, Kalman filter	[80]	1–4.5 m/4 humans/ECG	Error rate: 2.25%–4.6%
RR	FFT	[81]	70 cm/1 humans/manual	SNR (stationary): 20 dB SNR (speaking): 15 dB SNR (handwriting): 16 dB
RR, HR	FFT, Continuous Wavelet Transform (CWT)	[82]	5.4 m/8 humans/ECG	Success rate: RR: 94% HR: 89%
RR	FFT, distance deviation threshold (for random body motion detection)	[83]	1 m/22-year-old male/microphone	Normal breathing, apnea, macro-motion detection