Sensor device and method for measuring and determining a pulse arrival time (PAT) value

1. A method for measuring and determining a single pulse arrival time (PAT) value at a given location of a user, wherein the PAT value is the time for a pulse to arrive at said location following a contraction of the heart of the user, using a sensor device comprising a photoplethysmographic (PPG) multichannel sensor formed from a plurality of PPG sensor channels and being adapted to measure a set of PPG signals, each PPG signal being measured by one of the PPG sensor channels when the multichannel PPG sensor is in contact with the user; comprising:

measuring noninvasively said set of PPG signals;

extracting a plurality of features from each of the measured PPG signals;

selecting a subset from the set of PPG signals based on the extracted features; and

processing the selected subset of PPG signals to noninvasively determine the PAT value,

wherein said selecting the subset of PPG signals comprises:

projecting the plurality of features extracted from each of the measured PPG signal into a set of points in a N-dimensional feature space where N corresponds to the number of extracted features for each PPG signal;

clustering the set of points according to a distance criterion; and

selecting points being located at a representative cluster, the selected points corresponding to the subset of PPG signals.

2. The method according to claim 1 , wherein said extracting the plurality of features comprises fitting a parametric model to each PPG signal of said set of PPG signals, the features corresponding to the parameters of the fitted model.

3. The method according to claim 1 , wherein said processing the selected subset of PPG signals comprises estimating a representative value of at least one of the features extracted from each of the selected subset of PPG signals.

4. The method according to claim 1 , wherein said clustering the set of points comprises:

calculating a representative point;

calculating a distance for the set of points to the representative point; and

clustering together those points being located at a distance smaller than a distance threshold.

5. The method according to claim 4 , wherein said calculating the representative point comprises calculating the median point of the set of points.

6. The method according to claim 4 , wherein said calculating the distance for the set of points comprises calculating a Mahalanobis distance.

7. The method according to claim 4 , wherein said distance threshold is calculated from a histogram of the distances for the set of points to the representative point.

8. The method according to claim 1 , wherein the sensor device further comprises an electrocardiography (ECG) sensor for measuring an ECG signal, an impedance cardiography (ICG) sensor for measuring an ICG signal, and a phonocardiography (PCG) sensor for measuring a PCG signal; further comprising:

measuring the ECG signal with the ECG sensor such as to detect a R-Wave corresponding to the onset of left-ventricular depolarization,

triggering the PCG signal measured by the PCG sensor and the ICG signal measured by the ICG sensor by the detected R-Wave;

performing an ensemble average of the triggered PCG signal and ICG signal;

calculating the maximum of the envelope of the ensemble-averaged PCG signal; and

detecting a maximum of the third derivative of the ensemble-averaged ICG signal being closest to the maximum of the envelope such as to determine a value of a pre-ejection period (PEP).

9. The method according to claim 8 , wherein the determined value of the pre-ejection period (PEP) is being used in combination with the determined PAT value to estimate a pulse transit time (PTT) value.

10. The method according to claim 9 , further comprising calculating a blood pressure value from the estimated PTT value by measuring reference blood pressure values using a brachial cuff, and calibrating PTT values according to measured reference blood pressure values.

11. The method according to claim 9 , wherein further comprising estimating a pulse wave velocity (PWV) value by dividing the estimated PTT value by a measured distance of the user's body.

12. The method according to claim 11 , further comprising calculating a blood pressure value from the estimated PWV value by measuring reference blood pressure values using a brachial cuff, and calibrating PWV values according to measured reference blood pressure values.

13. A sensor device for measuring and determining noninvasively a pulse arrival time (PAT) value at a given location of a user, wherein the PAT value is the time for a pulse to arrive at said location following a contraction of the heart of the user, the sensor device comprising a photoplethysmographic (PPG) multichannel sensor formed from a plurality of PPG sensor channels and being adapted to measure a set of PPG signals, each PPG signal being measured by one of the PPG sensor channels when the multichannel PPG sensor is in contact with the user, wherein the sensor device is configured to determine the PAT value by:

measuring noninvasively said set of PPG signals;

extracting a plurality of features from each of the measured PPG signals;

selecting a subset from the set of PPG signals based on the extracted features; and

processing the selected subset of PPG signals to noninvasively determine the PAT value,

wherein said selecting the subset of PPG signals comprises:

projecting the plurality of features extracted from each of the measured PPG signal into a set of points in a N-dimensional feature space where N corresponds to the number of extracted features for each PPG signal;

clustering the set of points according to a distance criterion; and

selecting points being located at a representative cluster, the selected points corresponding to the subset of PPG signals.

14. The sensor device according to claim 13 , further comprising an impedance cardiography (ICG) sensor for measuring an ICG signal, an electrocardiography (ECG) sensor for measuring an ECG signal, and a phonocardiography (PCG) sensor for measuring a PCG signal; the ICG sensor, ECG sensor, and PCG sensor being used in combination for determining a value of a pre-ejection period (PEP).

15. The sensor device according to claim 14 , wherein the sensor device is further configured to determine the value of the PEP by:

measuring the ECG signal with the ECG sensor such as to detect a R-Wave corresponding to the onset of left-ventricular depolarization,

triggering the PCG signal measured by the PCG sensor and the ICG signal measured by the ICG sensor by the detected R-Wave;

performing an ensemble average of the triggered PCG signal and ICG signal;

calculating the maximum of the envelope of the ensemble-averaged PCG signal; and

detecting a maximum of the third derivative of the ensemble-averaged ICG signal being closest to the maximum of the envelope such as to determine the value of PEP.

16. A non-transitory computer readable medium comprising program code portions to be executed by a signal processing device in order to carry out a method for measuring and determining noninvasively a single pulse arrival time (PAT) value of a user, wherein the PAT value is the time for a pulse to arrive at a given location following a contraction of the heart of the user, using a sensor device comprising a photoplethysmographic (PPG) multichannel sensor formed from a plurality of PPG sensor channels and being adapted to measure a set of PPG signals, each PPG signal being measured by one of the PPG sensor channels when the multichannel PPG sensor is in contact with the user; the method comprising:

measuring noninvasively said set of PPG signals;

extracting a plurality of features from each of the measured PPG signals;

selecting a subset from the set of PPG signals based on the extracted features; and

processing the selected subset of PPG signals to noninvasively determine the PAT value;

when said program is executed by said signal processing device,

wherein said selecting the subset of PPG signals comprises:

projecting the plurality of features extracted from each of the measured PPG signal into a set of points in a N-dimensional feature space where N corresponds to the number of extracted features for each PPG signal;

clustering the set of points according to a distance criterion; and

selecting points being located at a representative cluster, the selected points corresponding to the subset of PPG signals.