Table 1.

Accuracy of previous blood pressure estimations based on PTT, PAT, and other methods.

	Systolic Blood Pressure Errors (Mean ± Standard Deviation)	Diastolic Blood Pressure Errors (Mean ± Standard Deviation)	Operating Range Systolic Blood Pressure	Operating Range Diastolic Blood Pressure
PTT-based Methods
Two-step algorithm developed by machine learning [17]	0.07 ± 7.1 mmHg	−0.08 ± 6.0 mmHg	Reduced accuracy for hypotension	Reduced accuracy for hypotension
B.P. estimation based on PTT and photoplethysmography intensity ratio (PIR) [18]	−0.37 ± 5.21 mmHg	−0.08 ± 4.06 mmHg	Reduced accuracy for hypertension	Reduced accuracy for hypertension
B.P. estimation based on PTT and intensity ratio of the first derivative wave of PPG (1st-dPIR) [19]	2.88 ± 7.75 mmHg	2.80 ± 4.38 mmHg
Proceeding PTT-based method on the repeatability test [20]	0.0 ± 5.3 mmHg	0.0 ± 2.9 mmHg	80–150 mmHg	60–120 mmHg
Proceeding PTT-based method using regression coefficients [20]	1.4 ± 10.2 mmHg	2.1 ± 7.3 mmHg	80–150 mmHg	60–120 mmHg
PAT-based Methods
Estimating beat-by-beat blood pressure using Chen’s method [21]	−0.5 ± 5.3 mmHg	4.1 ± 3.4 mmHg
Standard pulse arrival time based method calculations [22]	0 ± 3 mmHg	0 ± 3 mmHg
Using a linear correlation of systolic blood pressure and a non-linear correlation of diastolic blood pressure and PAT [23]	0.2 ± 5.8 mmHg	0.4 ± 5.7 mmHg
Model-driven method: Logarithmic [26]	−0.512 ± 8.793 mmHg	−0.148 ± 3.622 mmHg
Model-driven method: Inverse [26]	−0.008 ± 8.203 mmHg	−0.078 ± 3.448 mmHg
Model-driven method: Inverse Square [26]	−0.358 ± 8.084 mmHg	−0.066 ± 3.574 mmHg
Other Methods
Estimating blood pressure based on pulse morphology of PPG [24]	0.043 ± 5.001 mmHg	0.011 ± 3.689 mmHg
Blood pressure prediction based on demographic and physiological partitioning [25]	Mean absolute error = 6.9 mmHg	Mean absolute error = 5 mmHg	80–220 mmHg	45–120 mmHg